EO4SD-Urban Project: Arusha City Report...Applications (EO4SD-Urban) to the World Bank Urban...

Earth Observation for Sustainable Development

Urban Development Project

This project has received funding from the European Union’s Horizon 2020 research and innovation programme under grant agreement No 685761.

ESA Ref: AO/1-8346/15/I-NB

Doc. No.: City Operations Report

Issue/Rev.: 3.0

Date: 24.04.2018

EO4SD-Urban Project: Arusha City Report

Lead: Partners: Financed by:

Earth Observation for Sustainable Doc. No.: City-Operations Report

Development – Urban Project Issue/Rev-No.: 3.0

EO4SD-Urban Arusha City Operations Report Page I

Consortium Partners

No. Name Short Name Country

1 GAF AG GAF Germany

2 Système d'Information à Référence Spatiale SAS SIRS France

3 GISAT S.R.O. GISAT Czech Republic

4 Egis SA EGIS France

5 Deutsche Luft- und Raumfahrt e. V DLR Germany

6 Netherlands Geomatics & Earth Observation B.V. NEO The Netherlands

7 JOANNEUM Research Forschungsgesellschaft mbH JR Austria

8 GISBOX SRL GISBOX Romania

Disclaimer:

The contents of this document are the copyright of GAF AG and Partners. It is released by GAF AG on

the condition that it will not be copied in whole, in section or otherwise reproduced (whether by

photographic, reprographic or any other method) and that the contents thereof shall not be divulged to

any other person other than of the addressed (save to the other authorised officers of their organisation

having a need to know such contents, for the purpose of which disclosure is made by GAF AG) without

prior consent of GAF AG.



EO4SD-Urban Arusha City Operations Report Page II

Summary

This document contains information related to the provision of geo-spatial products from the European

Space Agency (ESA) supported project “Earth Observation for Sustainable Development” Urban Applications (EO4SD-Urban) to the World Bank Urban Planning Study for Tanzania programme for

the City of Arusha.

Affiliation/Function Name Date

Prepared GAF AG J. Freitas Santos, D.

Angelova,

18/04/2018

Reviewed GAF AG A. Broszeit 19/04/2018

Approved GAF AG, Project Coordinator T. Haeusler 24/04/2018

The document is accepted under the assumption that all verification activities were carried out correctly

and any discrepancies are documented properly.

Distribution

Affiliation Name Copies

ESA Z. Bartalis electronic copy

World Bank Chyi-Yun Huang electronic copy

Document Status Sheet

Issue Date Details

1.0 18/08/2017 First Document Issue

2.0 15/11/2017 Second Document Issue

3.0 24/04/2018 Third Document Issue

Document Change Record

# Date Request Location Details

1 15/11/2017 Ch. 2.6, 2.7,

2.8, 3.5, 3.6,

4.4, 4.5, 4.6

and Annexe 2

Three products were added to the Service

Operations Report: 1. Urban Green Areas 2.

Planned and Unplanned Settlement Areas 3.

Population Distribution and Density

2 15/11/2017 Annexe 1 Description of processing methods included

for: Transport Network, Urban Green Areas,



EO4SD-Urban Arusha City Operations Report Page III

Planned and Unplanned Settlement Areas,

Population Distribution and Density.

3 24/04/2018 Ch. 2.1, 3.3.1,

4.2, 4.3 and

4.4

After a User’s request, the spatial datasets for

the LULC, Transport Network and Green

Areas products were adjusted. These updates

implied changes in the following Sections of

the City Operations Report.



EO4SD-Urban Arusha City Operations Report Page IV

Executive Summary

The European Space Agency (ESA) has been working closely together with the International Finance

Institutes (IFIs) and their client countries to demonstrate the benefits of Earth Observation (EO) in the

IFI development programmes. Earth Observation for Sustainable Development (EO4SD) is a new ESA

initiative, which aims to achieve an increase in the uptake of satellite based information in the regional

and global IFI programmes. The overall aim of the EO4SD Urban project is to integrate the application

of satellite data for urban development programmes being implemented by the IFIs or Multi-Lateral

Development Banks (MDBs) with the developing countries. The overall goal will be achieved via

implementation of the following main objectives:

To provide a service portfolio of Baseline and Derived urban-related geo-spatial products

To provide the geo-spatial products and services on a geographical regional basis

To ensure that the products and services are user-driven

This Report describes the generation and the provision of EO-based information products to the World

Bank (WB) supported Technical Assistance project “Urban Planning Study-Impact & Effectiveness

of Urban Planning on City Spatial Development” for Tanzania and the counterpart City authorities

in Arusha. The Report provides a Service Description by referring to the user driven service

requirements and the associated product list with the detailed product specifications. The following

products were requested:

Urban Land Use/ Land Cover

Urban Extent

Urban Green Areas

Extent and Type of Informal Settlements


Transport Infrastructure - Road Network

The current Version of this Report contains the description of the generation and delivery of the Land

Use/Land Cover (LU/LC) and the LU/LC Change between two time periods of 2005 and 2015 as well

as the Transport Infrastructure product. The Derived products of Green Areas, Informal Settlements and

Population Density product are also included in this Report. Furthermore, it focuses on the

improvements of the LULC, Transport infrastructure and Green Areas products which have been done

based on the User Utility Review.

This City Operations Report for Arusha systematically reviews the main production steps involved and

importantly highlights the Quality Control (QC) mechanisms involved; the steps of QC and the

assessment of quality is provided in related QC forms in the Annexe of this Report. There is also the

provision of standard analytical work undertaken with the products which can be further included as

inputs into further urban development assessments, modelling and reports.



EO4SD-Urban Arusha City Operations Report Page V

This page is intentionally left blank!



EO4SD-Urban Arusha City Operations Report Page VI

Table of Contents

1 GENERAL BACKGROUND OF EO4SD-URBAN ................................................................... 1 2 SERVICE DESCRIPTION........................................................................................................... 1

2.1 STAKEHOLDERS AND REQUIREMENTS ..................................................................................... 1

2.2 SERVICE AREA SPECIFICATION ............................................................................................... 2

2.3 PRODUCT LIST AND PRODUCT SPECIFICATIONS ...................................................................... 3

2.4 LAND USE/LAND COVER NOMENCLATURE ............................................................................. 3

2.5 TRANSPORT INFRASTRUCTURE - ROAD NETWORK ................................................................. 6

2.6 URBAN GREEN AREAS ............................................................................................................. 6

2.7 PLANNED AND UNPLANNED SETTLEMENT AREAS .................................................................. 7

2.8 POPULATION DISTRIBUTION AND DENSITY ............................................................................. 7

2.9 TERMS OF ACCESS ................................................................................................................... 8

3 SERVICE OPERATIONS ............................................................................................................ 9

3.1 SOURCE DATA ......................................................................................................................... 9

3.2 PROCESSING METHODS ......................................................................................................... 10

3.3 ACCURACY ASSESSMENT OF MAP PRODUCTS ...................................................................... 10

3.3.1 The Applied Sampling Design ........................................................................................................ 10

3.3.2 The Applied Response Design ........................................................................................................ 12

3.3.3 The Applied Analysis Design .......................................................................................................... 13

3.4 ACCURACY ASSESSMENT OF TRANSPORT NETWORK ........................................................... 14

3.5 ACCURACY ASSESSMENT OF URBAN GREEN AREAS ............................................................ 16

3.6 ACCURACY ASSESSMENT OF PLANNED AND UNPLANNED SETTLEMENTS ........................... 17

3.7 QUALITY CONTROL/ASSURANCE .......................................................................................... 19

3.8 METADATA ............................................................................................................................ 20

4 ANALYSIS OF MAPPING RESULTS ..................................................................................... 22

4.1 URBAN EXTENT – DEVELOPMENTS 2000, 2005, 2010 AND 2015 .......................................... 22

4.2 LAND COVER LAND USE 2005 AND 2015 .............................................................................. 24

4.2.1 Spatial Distribution of Main LU/LC Change Categories ................................................................ 26

4.2.2 Changes of Agricultural Areas ........................................................................................................ 28

4.3 TRANSPORT NETWORK .......................................................................................................... 30

4.4 URBAN GREEN AREAS ........................................................................................................... 30

4.5 PLANNED AND UNPLANNED SETTLEMENT AREAS ................................................................ 32

4.6 POPULATION DISTRIBUTION AND DENSITY ........................................................................... 36

4.7 CONCLUDING POINTS ............................................................................................................ 38

5 REFERENCES ............................................................................................................................ 39



EO4SD-Urban Arusha City Operations Report Page VII

Annexes

Annex 1: Processing Methods for EO4SD-Urban Products

Annex 2: Filled Quality Control Sheets

List of Figures

Figure 1: Illustration of Core and Peri-Urban Areas of Mapping for Arusha. ................................... 2

Figure 2: Mapping result of the city of Arusha of the year 2015 overlaid with randomly distributed

sample points used for accuracy assessment..................................................................... 13

Figure 3: Example of the applied sampling design to generate randomly distributed point for the

Accuracy Assessment of the road network. ...................................................................... 14

Figure 4: Secondary sampling grid to generate the sampling points at spatial intersection of roads

and grid cells. Roads are represented as white lines, grid as black lines and final sampling

point as black dots. ............................................................................................................ 15

Figure 5: Result of the Urban Green Area mapping in Arusha (change product) with sampling points

used for product validation. .............................................................................................. 16

Figure 6: Result of the Informal Settlement Area mapping in Arusha (change product) with sampling

points used for product validation. .................................................................................... 18

Figure 7: Quality Control process for EO4SD-Urban product generation. At each intermediate

processing step output properties are compared against pre-defined requirements. ......... 19

Figure 8: Urban Extent developments in the epochs 2000 to 2005, 2005 to 2010 and 2010 to 2015 in

Arusha and surrounding region. ........................................................................................ 23

Figure 9: Urban Extent developments in the epochs 2000 to 2005, 2005 to 2010 and 2010 to 2015 in

Arusha within the Core Urban Area. ................................................................................ 23

Figure 10: Left: Overview of the Service Area. Right: Detailed Land Cover Land Use 2015 in Arusha.

.......................................................................................................................................... 24

Figure 11: Detailed Land Cover Land Use 2005 structure: Presented as Overall, Core Urban and in

Peri-Urban in % (left) and km2 (right). ............................................................................ 25

Figure 12: Detailed Land Cover Land Use 2015 structure: Presented as Overall, Core Urban and in

Peri-Urban in % (left) and km2 (right). ............................................................................. 25

Figure 13: Land Cover Land Use Change Types - Spatial Distribution. ........................................... 27

Figure 14: Land Cover Land Use Change Types 2000-2015 - overall, in Core Urban and in Peri-Urban

Zone in % (left) and km² (right) in Arusha. ...................................................................... 27

Figure 15: Spatial distribution of changes from Agricultural Areas to other Classes between 2005 and

2015. ................................................................................................................................. 28

Figure 16: Changes of Agricultural Areas into other LU classes between 2005 and 2015; Presented as

Overall, Core Urban and in Peri-Urban Zone in % (left) and km2 (right). ....................... 29

Figure 17: Transport Network of Arusha in 2005 and 2015. ............................................................. 30

Figure 18: Map overview of Urban Green Areas in Arusha. Green area loss, gain and stable green

areas can be identified. The map is delivered as separate product of high resolution for

printing at paper size DIN A0, which is 84.1 cm x 118.9 cm. .......................................... 31

Figure 19: Percentage of urban green areas within the core area of Arusha. The pie chart illustrates the

status and change of urban green areas in-between 2005 and 2015. ................................ 31

Figure 20: Bar charts for both points in time presenting the total area of urban greenery versus non-

green areas. ....................................................................................................................... 32

Figure 21: Planned settlement areas in Arusha in 2015. .................................................................... 33

Figure 22: Unplanned settlement areas in Arusha in 2015. ................................................................ 33

Figure 23: Map overview of changes in planned and unplanned settlement areas in Arusha during the

years 2005 and 2015. ........................................................................................................ 34

Figure 24: Percentage of planned and unplanned areas within the core area of Arusha. The pie chart

illustrates the status and change of planned and unplanned areas in-between 2005 and 2015.

…………………………………………………...…………………………………….35

Figure 25: Bar charts for both points in time presenting the total area of planned and unplanned

settlement areas. ................................................................................................................ 35



EO4SD-Urban Arusha City Operations Report Page VIII

Figure 26: Overview Map of Population Distribution Change in Arusha (2005 – 2015). ................. 36

Figure 27: Population Distribution Change within the Core Urban Districts of Arusha between 2005

and 2015. ........................................................................................................................... 37

Figure 28: Changes in Population Distribution, in relation to soil sealing degree in Arusha between

2005 and 2015. .................................................................................................................. 37

List of Tables

Table 1: LU/LC Nomenclature for 2005 and 2015............................................................................ 4 Table 2: Nomenclature used for the mapping and identification of Urban Green Areas. ................. 7

Table 3: Nomenclature used for the mapping and identification of Planned and Unplanned

Settlement Areas. ................................................................................................................. 7

Table 4: Nomenclature used for the mapping and identification of Population Distribution and

Density. ................................................................................................................................ 8 Table 5: Number of sampling points for the EO4SD-Urban mapping classes after applied sampling

design with information on overall land cover by class. ................................................... 12 Table 6: Validation result of the complemented Transport Network in Arusha, which is based on

OSM data. .......................................................................................................................... 15 Table 7: Calculation of the minimum number of samples according Goodchild et al. (1994). ....... 16

Table 8: Results of the Accuracy Assessment of Urban Green Areas in Arusha, 2005. ................. 17 Table 9: Results of the Accuracy Assessment of Urban Green Areas in Arusha, 2015. ................. 17 Table 10: Calculation of the minimum number of samples according Goodchild et al. (1994). ....... 18

Table 11: Results of the Accuracy Assessment of Informal Settlement Areas in Arusha, 2015. ...... 19 Table 12: Detailed information on area and percentage of total area for each class for 2005 and 2015

as well as the changes. ....................................................................................................... 26 Table 13: Overall LU/LC Statistics. .................................................................................................. 28 Table 14: Statistics of changes of Agricultural areas. ....................................................................... 29



EO4SD-Urban Arusha City Operations Report Page IX

List of Abbreviations

CDS City Development Strategy

CS Client States

DEM Digital Elevation Model

DLR German Space Agency

EEA European Environmental Agency

EGIS Consulting Company for Environmental Impact Assessment and Urban Planning, France

EO Earth Observation

ESA European Space Agency

EU European Union

GAF GAF AG, Geospatial Service Provider, Germany

GIS Geographic Information System

GISAT Geospatial Service Provider, Czech Republic

GISBOX Romanian company with activities of Photogrammetry and GIS

GUF Global Urban Footprint

HR High Resolution

HRL High Resolution Layer

IFI International Financing Institute

INSPIRE Infrastructure for Spatial Information in the European Community

ISO/TC 211 Standardization of Digital Geographic Information

JR JOANNEUM Research, Austria

LC / LU Land Cover / Land Use

LULCC Land Use and Land Cover Change

MMU Minimum Mapping Unit

NDVI Normalized Difference Vegetation Index

NEO Geospatial Service Provider, The Netherlands

QA Quality Assurance

QC Quality Control

QM Quality Management

SP Service Provider

VHR Very High Resolution

WB World Bank

WBG World Bank Group



EO4SD-Urban Arusha City Operations Report Page 1

1 General Background of EO4SD-Urban

Since 2008 the European Space Agency (ESA) has worked closely together with the International

Finance Institutes (IFIs) and their client countries to harness the benefits of Earth Observation (EO) in

their operations and resources management. Earth Observation for Sustainable Development (EO4SD)

is a new ESA initiative, which aims to achieve an increase in the uptake of satellite based information

in the regional and global IFI programmes. The EO4SD-Urban project initiated in May 2016 (with a

duration of 3 years) has the overall aim to integrate the application of satellite data for urban

development programmes being implemented by the IFIs with the developing countries. The overall

goal will be achieved via implementation of the following main objectives:

To provide the services on a regional basis (i.e. large geographical areas); in the context of the

current proposal with a focus on S. Asia, SE Asia and Africa, for at least 35-40 cities.

To ensure that the products and services are user-driven; i.e. priority products and services to

be agreed on with the MDBs in relation to their regional programs and furthermore to implement

the project with a strong stakeholder engagement especially in context with the validation of the

products/services on their utility.

To provide a service portfolio of Baseline and Derived urban-related geo-spatial products that

have clear technical specifications, and are produced on an operational manner that are

stringently quality controlled and validated by the user community.

To provide a technology transfer component in the project via capacity building exercises in the

different regions in close co-operation with the MDB programmes.

This Report supports the fulfilment of the third objective which requires the provision of geo-spatial

Baseline and Derived geo-spatial products to various stakeholders in the IFIs and counterpart City

authorities. The Report provides a Service Description, and then in Chapter 3 systematically reviews the

main production steps involved and importantly highlights whenever there are Quality control (QC)

mechanisms involved with the related QC forms in the Annexe of this Report. The description of the

processes is kept intentionally at a top leave and avoiding technical details as the Report is considered

mainly for non-technical IFI staff and experts and City authorities. Finally Chapter 4 presents the

standard analytical work undertaken with the products which can be inputs into further urban

development assessments, modelling and reports.

2 Service Description

The following Section summarises the service as it has been realised for the city of Arusha, Tanzania

within the EO4SD-Urban Project and as it had been delivered to the Task Team Leader (TTL) of the

World Bank programme Urban Planning Study for Tanzania in August 2017.

2.1 Stakeholders and Requirements

The EO4SD-Urban products described in this Report were provided to the World Bank (WB) Technical

Assistance project “Urban Planning Study-Impact & Effectiveness of Urban Planning on City

Spatial Development” for Tanzania. The Urban Planning Study is linked to the larger WB project

‘Tanzania Strategic Cities Project’ (TSCP). The TSCP supports the seven secondary cities: Tanga,

Arusha, Mwanza, Kigoma, Dodoma, Mbeya and Mtwara, in addition to the Capital Development

Authority (CDA) in Dodoma, the national capital. These cities are of strategic importance as they are

within the top ten most populous cities in Tanzania with high population growth rates. In the context of

the current EO4SD Urban project, three of the overall seven cities - Arusha, Kigoma, Dodoma - could

be provided with the geo-spatial datasets. The higher objective of the Urban Planning Study, as stated

in its Draft Concept Note (World Bank, 2016a) is, “to enhance the urban development and inform policies and development strategies of cities in Tanzania through gaining further insights on urban

planning system and development processes and the effectiveness of master and detailed urban




plans.” It aims to increase the appreciation of practitioners and local authorities on the importance of urban plans for guiding urban growth and mitigating the potential problems and higher costs of

retrofitting unplanned development. In this context, the Bank is collaborating with the Consortium to

investigate the spatial development characteristics of Tanzanian cities with and without urban plans, and

attempt to assess the impact and effectiveness of such urban plans. This specific Version of the Report

focusses on providing the new results from the improved LULC, Transport infrastructure and Green

areas products.

2.2 Service Area Specification

The Areas of Interest (AOI) for mapping the Urban Area and the Peri-Urban Areas for Arusha was

depicted in a power point slide, and sent to the Users for verification. The boundaries depicted were

based on the municipality and administration boundaries of the cities. These boundaries were obtained

from the GADM database of Global Administrative Areas (http://www.gadm.org/).

As the city administrative boundary data availability is different for each country/city, the AoIs for the

Urban and Peri-Urban Area were in some cases adjusted to areas which could provide the best examples

of the geo-spatial products that Users may require. The adjustments are based on population distribution

data from LandScan (http://web.ornl.gov/sci/landscan/) and on visual interpretation of built-up areas as

evidenced on Google Earth. LandScan is the finest resolution global population distribution data (~1km

spatial resolution) available and represents an ambient population.

Figure 1: Illustration of Core and Peri-Urban Areas of Mapping for Arusha.

The Core region has an area of 211 km2 and the Peri-Urban has an area of 414 km2, for a total service

or mapping area of 625 km2.




2.3 Product List and Product Specifications

During the discussions related to the AoIs the potential geo-spatial products that could be provided for

the Cities were also reviewed with the WB Team and users. It was noted that the Baseline Land

Use/Land Cover (LU/LC) products (for the Core and Peri-Urban areas) were a standard product that

would be provided for all Cities as it is required for the derived products. In the case of Arusha, the full

list of products for both the Core and Peri-Urban areas are as follows:

Urban Land Use/ Land Cover

Urban Extent

Urban Green Areas

Extent and Type of Informal Settlements


Transport Infrastructure - Road Network

For each of these products two time slots were used to provide historic and recent information; thus for

Arusha it was 2005 and 2015. Due to EO data availability the time epochs varied with a plus/minus year

or two around these 2 time slots. The current Report will focus on the provision of the Baseline LU/LC

products and the Transport Infrastructure for Arusha.

2.4 Land Use/Land Cover Nomenclature

A pre-cursor to starting production was the establishment with the stakeholders on the relevant Land

Use/Land Cover (LU/LC) nomenclature as well as class definitions. The approach taken was to use a

standard remote sensing based LU/LC nomenclature and then adapt it to the user’s LU requirements.

Thus the remote-sensing based LU/LC classes in the urban context can be grouped into 5 Level 1 classes,

which are Artificial Areas, Natural/ Semi Natural, Agricultural, Wetland, Water Bodies. These classes

can then be sub-divided into several different more detailed classes such that the dis-aggregation can be

down to Level 2-4. This hierarchical classification system is often used in operational Urban mapping

programmes and is the basis for example of the European Commission’s Urban Atlas programme which

provides pan-European comparable LU/LC data for with regular updates. A depiction of the way the

levels and classes are structured is presented as follows:

Level I Artificial surfaces

- Level II Urban Fabric

Level III

Continuous Urban Fabric (Sealing Layer-S.L. > 80%)

Discontinuous Urban Fabric (S.L. 10% - 80%)

Discontinuous Dense Urban Fabric (S.L. 50% - 80%)

Discontinuous Medium Density Urban Fabric (S.L. 30% - 50%)

Discontinuous Low Density Urban Fabric (S.L. 10% - 30%)

Discontinuous Very Low Density Urban Fabric (S.L. < 10%)

- Level II

Industrial, commercial, public, military, private and transport units

Level III

Industrial, commercial, public, military and private units zoning data

Road and rail network and associated land (Open Street Map or In-country data needed)

- Level IV Fast transit roads and associated land

(Reference: European Union, 2011)




It should be noted that in the current project, the Level 1 classes were used as the basis for classification

of the Peri-Urban areas using the High Resolution (HR) data such as Landsat or Sentinel. However, for

the Core Urban areas using the Very High Resolution (VHR) data it was possible to go down to Level

III and IV. The different Levels, classes and sub-classes from the remote sensing based Urban

classification, were adapted to the User requirements based on existing Master Plans for cities and/or

direct advise from the User on critical classes required. The final LU/LC nomenclature had to be

endorsed by the user before productions started.

In order to link the Urban Atlas classes described in the previous Section with the Tanzanian cities, the

Consortium used the documentation provided by the Cities for their Master Plans, and assessed which

LU/LC classes could be mapped with remote sensing and linked to the Urban Atlas nomenclature. The

merging of LU/LC classes was provided to the WB Team and users for review and endorsement. For

Arusha, the Consortium referred to the Arusha Master Plan 2015-2035 (2016), Ministry of Lands,

Housing and Human Development, Tanzania in order to assess which classes could be mapped with

remote sensing. See Table 1 for final LU nomenclature used for the different epochs.

Table 1: LU/LC Nomenclature for 2005 and 2015.

2005 2015

Level I Level II Level III Level IV Level I Level II Level III Level IV

1000

Artificial

Surfaces

1100

Residential

1110 Very Low

Density

1000

Artificial

Surfaces

1100

Residential

1110 Very

Low Density

1120 Low

Density

1120 Low

Density

1130 Medium

Density

1130 Medium

Density

1140 High

Density

1140 High

Density

1150 Very High

Density

1150 Very

High Density

1200

Industrial,

Commercia

l, Public,

Military,

Private and

Transport

Units

1210 Industrial,

Commercial,

Public, Military

and Private

Units

- 1200

Industrial,

Commercial,

Public,

Military,

Private and

Transport

Units

1210

Industrial,

Commercial,

Public,

Military and

Private Units

1211 Commercial

- 1212 Industry

- 1213 University

- 1214 Schools

- 1215 Government

- 1216 Military

- 1217 Hospitals

- 1218

Public Buildings

1220 Roads 1221

Arterial

1220

Roads

1221 Arterial

1222

Collector

1222 Collector

1230 Railway 1230 Railway

1240 Airport 1240 Airport




1250 Port 1250 Port

1300 Mine,

Dump and

Constructio

n Sites

1310 Mineral

Extraction and

Dump Sites

1300 Mine,

Dump and

Construction

Sites

1310 Mineral

Extraction

and Dump

Sites

1320

Construction

Sites

1320

Construction

Sites

1330 Land

without Current

Use

1330

Land without

Current Use

1400 Urban

Open

Spaces

1410 Urban

Parks

1400

Urban Open

Spaces

1410

Urban Parks

1420 Recreation

Facilities (Sport

Facilities,

Stadiums, Golf

Courses, etc.)

1420

Recreation

Facilities

(Sport

Facilities,

Stadiums,

Golf Courses,

etc.)

1430

Cemeteries

1430

Cemeteries

2000.

Agricultu

ral Area

2000

Agricultu

ral Area

3000

Natural

and

Semi-

natural

Areas

3100 Forest

and Shrub

Lands

3000

Natural

and

Semi-

natural

Areas

3100

Forest and

Shrub Lands

3200

Natural

Areas

(Savannah,

Grassland)

3200

Natural Areas

(Savannah,

Grassland)

4000

Wetlands

4000

Wetlands

5000

Water

5000

Water

It is important to note that the possibility to classify at Level IV was highly dependent on the availability

of reliable Reference datasets from the City or sources such as Google Earth. This aspect is further

discussed in Chapter 3.




2.5 Transport Infrastructure - Road Network

The road hierarchy used in the classification is based on the international road classification standards;

this is for example defined by the European Commission (https://ec.europa.eu/transport/road_safety/-

specialist/knowledge/road/designing_for_road_function/road_classification_en).

Roads are divided into three groups-Arterial or through traffic flow routes (in our case Arterial Roads),

distributor road (in our case Collector Roads), and access roads (in our case Local Roads). The three

road types are defined as follows:

Arterial Roads:

Roads with a flow function allow efficient throughput of (long distance) motorized traffic. All

motorways and express roads as well as some urban ring roads have a flow function. The number of

access and exit points is limited. (https://ec.europa.eu/transport/road_safety/specialist/knowledge/-

road/designing_for_road_function/road_classification_en)

Collector Roads:

Roads with an area distributor function allow entering and leaving residential areas, recreational areas,

industrial zones, and rural settlements with scattered destinations. Junctions are for traffic exchange

(allowing changes in direction etc.); road sections between junctions should facilitate traffic in flowing.

(https://ec.europa.eu/transport/road_safety/specialist/knowledge/road/designing_for_road_function/roa

d_classification_en)

Local Roads:

Roads with an access function allow actual access to properties alongside a road or street. Both junctions

and the road sections between them are for traffic exchange. (https://ec.europa.eu/transport/-

road_safety/specialist/knowledge/road/designing_for_road_function/road_classification_en).

Arterial roads and collector roads were the main focus of the classification. These types of roads were

identified for the entire AoI. Within the geospatial dataset the road features can be identified within the

attribute table. A value of 1 is assigned to the arterial roads and a value of 2 to the collector lines.

Spatial Accuracy:

The Collector roads and Arterial roads are integrated within the LULC mapping by applying a buffer

around the road centre lines of 12.0 m for the arterial roads and 7.5m for collector roads. The 7.5m are

set as the maximum allowable difference of the mapped centre line in comparison to the location in the

VHR imagery.

2.6 Urban Green Areas

Urban Green Areas were separated within the Land Use/Land Cover product only in the land use classes

1411 Urban Parks (Parks, Gardens) and 1412 Cemeteries. However, all other urbanised land use classes,

like 1100 Residential, 1200 Industrial, Commercial, Public, Military, Private and Transport Units, and

1300 Mine, Dump and Construction Sites can also be partially vegetated. In order to get an overall

synopsis of all Green Areas within the urbanised classed, i.e. all sub-classes under 1000 Artificial

Surfaces, a further land cover classification of the VHR data was performed by applying a semi-

automated classification algorithm. Forested areas, Grassland and Agricultural areas which are part of

the LULC product were not further considered in this Urban Green Areas land cover. The applied

nomenclature and class coding is given in Table 2.




Table 2: Nomenclature used for the mapping and identification of Urban Green Areas.

Single date

Code 0 Non-green area

Code 1 Urban green area

Code 255 Non-urban areas. All areas that do not fall in the Level I class of the Land Use

Land Cover product (See Table 1).

Change product

Code 0 Non-green area. No vegetated surfaces occurring on “Artificial Surfaces”, Level I, at both points in time.

Code 1 Permanent urban green area. Vegetated surfaces in 2005 and 2015.

Code 2 Loss of urban green area. Vegetated areas in 2005, which changed to non-

vegetated areas in 2015.

Code 3 New urban green area. Non-vegetated surfaces in 2005 with vegetation cover in

2015.

Code 255 Non-Urban Areas. All Areas that do not fall in the Level I class of the Land Use

Land Cover product.

2.7 Planned and Unplanned Settlement Areas

For the classification of Planned and Unplanned Settlements only the Core Urban area was assessed.

Furthermore the LU classes that were targeted were only the residential areas, including the LULC

classes “1100 Residential” and “1211 Commercial Area.” It has to be noted that within the Commercial

Area class only those polygons with a residential component, like in the CBD were classified as planned

or unplanned settlements. The applied nomenclature is given in Table 3.

Table 3: Nomenclature used for the mapping and identification of Planned and Unplanned Settlement

Areas.

Single date

Planned settlements

Unplanned settlements

Change product

No change in planned settlement area

No change in unplanned settlement area

Expansion of planned settlement area

Expansion of unplanned settlement area

Decrease of planned settlement area

Decrease of unplanned settlement area

Unplanned to planned settlement area

2.8 Population Distribution and Density

The Population Distribution and Density product is calculated only for the residential areas (LULC class

11) within the Core Urban Districts of Arusha. The current product is derived from globally available

WorldPop data from 2015 with spatial resolution of 100m. Additionally, the population residuals are

adjusted, using interpolation of population counts from the official population census data. The historic

product is derived entirely based on interpolation of the official population census data. The applied

nomenclature is given in Table 4.




Table 4: Nomenclature used for the mapping and identification of Population Distribution and Density.

Single date

Population Density between 0 – 150 inhabitants/ km2








Population Density between 18001 - 41000 inhabitants/ km2

Change product

Unchanged Population Distribution

Up to -100% decrease

Up to 200% increase

201% - 400% increase



801% - 1000% increase

More than 1000% increase

2.9 Terms of Access

The Dissemination of the digital data and the Report was undertaken via FTP.




3 Service Operations

The following Sections present all steps of the service operations including the necessary input data, the

processing methods, the accuracy assessment and the Quality Control procedures. Methods are

presented in a top-level and standardised manner for all the EO4SD-Urban City Reports.

3.1 Source Data

This Section presents the remote sensing and ancillary datasets that were used. Different types of data

from several data providers have been acquired. A complete list of source data as well as a quality

assessment is provided in Annex 2.

A summary of the main data used is provided in the following Sections.

High Resolution Optical EO Data

The major data sources for the peri-urban current and historic mapping of urban LULUC, urban extent

and imperviousness were Landsat and Sentinel-2 data which were accessible and downloadable free of

charge.

Aster: as a source of historical data 1 scenes in total of Advanced Spaceborne Thermal Emission

and Reflection Radiometer Satellite (Aster) has been acquired available for the year 2006 that

is covering the whole area of interest. All image bands have been used as baseline data in the

production process. The ortho-rectification was achieved by an image to image registration

using the Global Landcover-Ortho image product. The NASA accuracy specification for the

Global Landcover-Ortho product was 50m RMSE, everywhere in the world.

Sentinel-2: the most recent data coverage comprises one Sentinel-2 data set. The data was

downloaded and processed at Level 1C.

Very High Resolution Optical EO Data

The VHR data for the core urban area mapping had to be acquired and purchased through commercial

EO Data Providers such as Airbus Defence and European Space Imaging.

It has to be noted that under the current collaboration project the VHR EO data had to be purchased

under mono-license agreements between GAF AG and the EO Data Providers. If EO data would have

to be distributed to other stakeholders then further licences for multiple users would have to be

purchased.

The following VHR sensor data have been acquired to cover the entire AoI:

WorldView-2 and GeoEye-1: o 6 scenes for 2016

o 2 scenes for 2015

Quickbird-2: o 6 scenes for 2005

Detailed lists of the used EO data as well as their quality is documented in the attached Quality Control

Sheets in Annex 2.

Ancillary Data

Data Provided by User:

o ARU_Landmarks.shp: Information on 70 objects that were used to assist the mapping

procedure. Mainly located around the city centre.

o Arusha_river.shp: Spatial location of rivers. Data was integrated in the LULC mapping

by using the spatial location of that data set. The polygons for the water class are based

on the VHR imagery.

o ARU_roads.shp: Similar to OSM data. A slight geometric shift in the data was

identified. Data was used to verify the Open Street Map (OSM) data, which was used

as basis for the road classes.




o tanzania_health_facility_registry6322wgs84.shp: 6323 Points of Interest (POIs) on

Health Centres, Hospitals and Dispensaries. The point based information was overlaid

on the LULC mapping, but no information was available on spatial precision and

completeness.

o updated_schools_november_2014_Dodoma.shp: Location of 12723 schools in

Tanzania. Data was used for map overlay as provided by User. No information was

available on spatial precision and completeness.

Open Street Map (OSM) data: OSM data is freely available and generated by volunteers across

the globe. The so called crowd sourced data is not always complete, but has for the most parts

of the world valuable spatial information. Data was downloaded to complement the Transport

Network layer and further enhanced. The spatial location of the OSM based streets was used a

geospatial reference.

Detailed lists of the used ancillary data as well as their quality is documented in the attached Quality

Control Sheets in Annex 2.

3.2 Processing Methods

Data processing starts at an initial stage with quality checks and verification of all incoming data. This

assessment is performed in order to guarantee the correctness of data before geometric or radiometric

pre-processing is continued. These checks follow defined procedures in order to detect anomalies,

artefacts and inconsistencies. Furthermore all image and statistical data were visualised and interpreted

by operators.

The main techniques and standards used for data analysis, processing and modelling for each product

are described in Annex 1.

3.3 Accuracy Assessment of Map Products

Data and maps derived from remote sensing contain - like any other map - uncertainties which can be

caused by many factors. The components, which might have an influence on the quality of the maps

derived from EO include quality and suitability of satellite data, interoperability of different sensors,

radiometric and geometric processing, cartographic and thematic standards, and image interpretation

procedures, post-processing of the map products and finally the availability and quality of reference

data. However, the accuracy of map products has a major impact on secondary products and its utility

and therefore an accuracy assessment was considered as a critical component of the entire production

and products delivery process. The main goal of the thematic accuracy assessment was to guarantee the

quality of the mapping products with reference to the accuracy thresholds set by the user requirements.

The applied accuracy assessments were based on the use of reference data, and applying statistical

sampling to deduce estimates of error in the classifications. In order to provide an efficient, reliable and

robust method to implement an accuracy assessment, there are three major components that had to be

defined: the sampling design, which determines the spatial location of the reference data, the response

design that describes how the reference data is obtained and an analyses design that defines the accuracy

estimates. These steps were undertaken in a harmonised manner for the validation of all the geo-spatial

products.

3.3.1 The Applied Sampling Design

The sampling design specifies the sample size, sample allocation and the reference assessment units

(i.e. pixels or image blocks). Generally, different sampling schemes can be used in collecting accuracy

assessment data including: simple random sampling, systematic sampling, stratified random sampling,

cluster sampling, and stratified systematic unaligned sampling. In the current project a single stage




stratified random sampling based on the method described by Olofson et al (20131) was applied which

used the map product as the basis for stratification. This ensured that all classes even very minor ones

were included in the sample.

However, in complex LU/LC products with many classes, this usually results in a large number of

strata (one stratum per LU/LC classes), of which some classes cover only very small areas (e.g. sport

fields, cemeteries) and not being adequately represented in the sampling. In order to achieve a

representative sampling for the statistical analyses of the mapping accuracy it was decided to extend

the single stage stratified random sampling. At the first stage the number of required samples was

allocated within each of the Level I strata (see Table 2). In the second stage all Level III classes that

were not covered by the first sampling, were grouped into one new stratum. Within that stratum the

same number of samples was randomly allocated as the Level I strata received. To avoid a clustering

of point samples within classes and to minimise the effect of spatial autocorrelation a minimum

distance in between the sample points was set to be 150 m. The final sample size for each class can be

considered to be as close as possible to the proportion of the area covered by each stratum considering

that the target was to determine the overall accuracy of the entire map.

The total sample size per stratum was determined by the expected standard error and the estimated

error rate based on the following formula which assumes a simple random sampling (i.e. the

stratification is not considered):

n = 𝑃∗𝐸𝑧 ²

n = number of samples per strata / map class

p = expected accuracy

q = 1 – p

E = Level of acceptable (allowable) sample error

Z = z-value (the given level of significance)

Hence, with an expected accuracy of p = 0.85, a 95% confidence level and an acceptable sampling

error of 5%, the minimum sample size is 196. A 10% oversampling was applied to compensate for

stratification inefficiencies and potentially inadequate samples (e.g. in case of cloudy or shady

reference data). For each Level I strata 215 samples have been randomly allocated. Afterwards, within

all classes of Level III (see Table 5) that did not received samples in the first run, additionally 215

samples were randomly drawn across all these classes.

1 Olofsson, P., Foody, G. M., Stehman, S. V., & Woodcock, C. E. (2013). Making better use of accuracy data in

land change studies: Estimating accuracy and area and quantifying uncertainty using stratified estimation.

Remote Sensing of Environment, 129, 122–131. doi:10.1016/j.rse.2012.10.031




Table 5: Number of sampling points for the EO4SD-Urban mapping classes after applied sampling

design with information on overall land cover by class.

Class Name Class

ID

No. of

Sampling

Points

Km²

coverage

Residential 1100 178 148.2

Commercial 1210 16 11.5

Roads 1220 150 5.8

Airport 1240 1 0.5

Mining 1310 66 0.9

Construction 1320 48 3.4

Land without current use 1330 7 4.2

Urban Parks 1410 12 0.2

Recreational Facilities 1420 1 0.9

Cemetery 1430 3 0.1

Agriculture 2000 215 359.2

Forest 3100 182 59

Natural Areas 3200 33 29.2

Water 5000 29 1.8

Total -- 941 624.8

3.3.2 The Applied Response Design

The response design determines the reference information for comparing the map labels to the reference

labels. Collecting reference data on the ground by means of intensive fieldwork is both costly and time

consuming and in most projects not feasible. The most cost effective reference data sources are VHR

satellite data with 0.5 m to 1 m spatial resolution. Czaplewski (2003)2 indicated that visual interpretation

of EO data is acceptable if the spatial resolution of EO data is sufficiently better compared to the

thematic classification system. However, if there are no EO data with better spatial resolution available,

the assessment results need to be checked against the imagery used in the production process.

The calculated number of necessary sampling points for each mapping category was randomly

distributed among the strata and overlaid to the VHR data of each epoch. The following Figure is

showing the mapping result with the overlaid sample points.

2 Czaplewski, R. L. (2003). Chapter 5: accuracy assessment of maps of forest condition: statistical design and

methodological considerations, pp. 115–140. In Michael A.Wulder, & Steven E. Franklin (Eds.), Remote

sensing of forest environments: concepts and case studies. Boston: Kluwer Academic Publishers (515 pp.).




Figure 2: Mapping result of the city of Arusha of the year 2015 overlaid with randomly distributed sample

points used for accuracy assessment.

In this way a reference information could be extracted for each sample point by visual interpretation of

the VHR data for all mapped classes. The size of the area to be observed had to be related to the

Minimum Mapping Unit (MMU) of the map product to be assessed. The reference information of each

sampling point was compared with the mapping results and the numbers of correctly and not-correctly

classified observations were recorded for each class. From this information the specific error matrices

and statistics were computed (see next Section).

3.3.3 The Applied Analysis Design

Each class usually has errors of both omission and commission, and in most situations, these errors for

a class are not equal. In order to calculate these errors as well as the uncertainties (confidence intervals)

for the area of each class a statistically sound accuracy assessment was implemented.

The confusion matrix is a common and effective way to represent quantitative errors in a categorical

map, especially for maps derived from remote sensing data. The matrices for each assessment epoch

were generated by comparing the “reference” information of the samples with their corresponding classes on the map. The Reference represented the “truth”, while the Map provided the data obtained

from the map result. Thematic accuracy for each class and overall accuracy is then presented in error

matrices (see Tables below). Unequal sampling intensity resulting from the random sampling approach

was accounted for by applying a weight factor (p) to each sample unit based on the ratio between the

number of samples and the size of the stratum considered3:

3 Selkowitz, D. J., & Stehman, S. V. (2011). Thematic accuracy of the National Land Cover Database (NLCD)

2001 land cover for Alaska. Remote Sensing of Environment, 115(6), 1401–1407. doi:10.1016/j.rse.2011.01.020.




�̂� = ( 1𝑀) ∑ 1𝜋 ℎ∗𝑥∈ ,

Where i and j are the columns and rows in the matrix, M is the total number of possible units (population)

and π is the sampling intensity for a given sample unit u in stratum h.

Overall accuracy and User and producer accuracy were computed for all thematic classes and 95%

confidence intervals were calculated for each accuracy metric.

The standard error of the error rate was calculated as follows: 𝜎ℎ = √ ℎ − ℎ𝑛ℎ where nh is the sample

size for stratum h and ph is the expected error rate. The standard error was calculated for each stratum

and an overall standard error was calculated based on the following formula: 𝜎 = √∑ 𝑤ℎ . 𝜎ℎ

In which 𝑤ℎ is the proportion of the total area covered by each stratum. The 95% Confidence Interval

(CI) is +/- 1.96*𝜎.

The confusion matrices are provided within the Annex 2 and showing the mapping error for each

relevant class. For each class the number of samples which are correctly and not correctly classified are

listed, which allows the calculation of the user and producer accuracies for each class as well as the

confidence interval at 95% confidence levels based on the formulae above.

The Land Use/Land Cover product for Arusha has an overall mapping accuracy of 88.8% with a

CI ranging from 86.8% to 90.9% at a 95% CI. The specific class accuracies are given in Annex 2.

3.4 Accuracy Assessment of Transport Network

The road network was partially integrated in the LULC map by selecting first and second level roads.

These are the Arterial Roads and the Collector Lines. For the accuracy assessment of the Road network

it should be noted that the sampling design, response design and analyses design are different from the

one used for validating the LULC maps. The Accuracy Assessment of the Transport Network is related

to the geospatial precision of the collected and digitised centerlines of the roads.

Sampling Methodology

A systematic random sampling was applied to define the primary and secondy sampling units. Over the

entire AoI a regular grid of 450m by 450m was created. Based on these grid cells a random selection of

2% sample cells were selected. An example is given in Figure 3 with the road network in grey, the grid

cells in black and the randomly selcted cells in green.

Figure 3: Example of the applied sampling design to generate randomly distributed point for the

Accuracy Assessment of the road network.




Within the randomly selected cells another grid of 150m distance was created (see Figure 4). All

intersection between the created road layer and the 150m grid were extracted as points. At all points the

road locations were visually checked and if any, the differences between spatial location on VHR

imagery and spatial location of the digitised lines recorded.

Figure 4: Secondary sampling grid to generate the sampling points at spatial intersection of roads and

grid cells. Roads are represented as white lines, grid as black lines and final sampling point as

black dots.

Overall 303 sampling points were created and their differences recorded. The result is presented in Table

6 as histogram of deviations. For the entire sampling population a Mean Difference of 1.66 m and a

Standard Deviation of 2.51 m was calculated.

Table 6: Validation result of the complemented Transport Network in Arusha, which is based on OSM

data.

Distance in m Frequency

0.0 177

1.5 5

3.0 50

4.5 35

6.0 14

7.5 12

Above 7.5 meters 10

By setting a maximum allowable difference of 7.5 m the distances are separated into two classes. Correct

street locations and in-correct street locations. The statistical analysis for the two classes reveal an

overall accuracy of 96.7% was achieved.




3.5 Accuracy Assessment of Urban Green Areas

The validation of the Green Area mapping results is done in a similar way as the validation for the Land

Use Land Cover product. The necessary amount of sampling points are calculated according to the

formula of Goodchild et al. (1994), which is given in Table 7.

Table 7: Calculation of the minimum number of samples according Goodchild et al. (1994).

Variables Values

p 0.85

q 0.15

E 0.05

z 1.96

196

n with 10% oversampling 215

with:

p = required accuracy of the data

q = 1-p


Z = value from table (for the given level of significance)

The calculated number of 215 sample points was randomly distributed among the entire map and

overlaid on the VHR data of each epoch. The following Figure shows the mapping result with the

overlaid sample points.

Figure 5: Result of the Urban Green Area mapping in Arusha (change product) with sampling points

used for product validation.




At each sample point location the reference data was collected by visual interpretation of the VHR data.

The size of the area to be observed had to be related to the Minimum Mapping Unit (MMU) of the map

product to be assessed. Finally, visual interpreted land cover type was compared with the mapping

results and the numbers of correctly and not-correctly classified observations were recorded. From this

information the specific error matrices and statistics were computed.

The confusion matrices show the mapping error for each relevant class. For each class the number of

samples which are correctly and not correctly classified are listed in the Tables below. They allow the

calculation of the user and producer accuracies for each class as well as the confidence interval at 95%

confidence levels based on the formulae above. The results of the Accuracy Assessment are listed in

Table 8 and Table 9 below, for 2005 and 2015 respectively.

Table 8: Results of the Accuracy Assessment of Urban Green Areas in Arusha, 2005.

Overall Accuracy 94.9 %.

Urban Green 2005 Reference Data

Totals 0 - Non-Urban Green Area 1 - Urban Green Area

0 - Non-Urban Green Area 97 9 106

1 - Urban Green Area 7 102 109

Totals 104 111 215

Table 9: Results of the Accuracy Assessment of Urban Green Areas in Arusha, 2015.


Urban Green 2015 Reference Data




Totals 128 87 215

The confusion matrices are additionally provided within the Quality Control documentation in Annex 2

and showing the mapping error for each relevant class. For each class the number of samples which are

correctly and not correctly classified are listed, which allows the calculation of the user and producer

accuracies for each class as well as the confidence interval at 95% confidence levels.

3.6 Accuracy Assessment of Planned and Unplanned Settlements

The validation of the Planned and Unplanned Settlement mapping results is done in a similar way as the

validation for the Land Use Land Cover and the Urban Green Area product. The necessary amount of

sampling points are calculated according to the formula of Goodchild et al. (1994), which is given in

Table 10.




Table 10: Calculation of the minimum number of samples according Goodchild et al. (1994).

Variables Values

p 0.85

q 0.15

E 0.05

z 1.96

196

n with 10% oversampling 215

with:

p = required accuracy of the data

q = 1-p


Z = value from table (for the given level of significance)

The calculated number of 215 sample points was randomly distributed among the entire map and

overlaid on the VHR data of each epoch. The following Figure shows the mapping result with the

overlaid sample points.

Figure 6: Result of the Informal Settlement Area mapping in Arusha (change product) with sampling

points used for product validation.

At each sample point location the reference data was collected by visual interpretation of the VHR data.

The size of the area to be observed had to be related to the Minimum Mapping Unit (MMU) of the map

product to be assessed. Finally, visual interpreted land cover type was compared with the mapping

results and the numbers of correctly and not-correctly classified observations were recorded. From this

information the specific error matrices and statistics were computed.




The confusion matrices show the mapping error for each relevant class. For each class the number of

samples which are correctly and not correctly classified are listed in the Tables below. They allow the

calculation of the user and producer accuracies for each class as well as the confidence interval at 95%

confidence levels based on the formulae above. The results of the Accuracy Assessment is listed in

Table 11 for 2015.

Table 11: Results of the Accuracy Assessment of Informal Settlement Areas in Arusha, 2015.


Settlements 2015 Reference Data

Totals Planned Settlement Area Unplanned Settlement Area

Planned Settlement Area 97 10 107

Unplanned Settlement Area 2 105 107

Totals 99 115 214

The confusion matrices are additionally provided within the Quality Control documentation in Annex 2

and showing the mapping error for each relevant class. For each class the number of samples which are

correctly and not correctly classified are listed, which allows the calculation of the user and producer

accuracies for each class as well as the confidence interval at 95% confidence levels.

3.7 Quality Control/Assurance

A detailed Quality Control and Quality Assurance (QC/QA) system has been developed which records

and documents all quality relevant processes ranging from the agreed product requirements, the different

types of input data and their quality as well as the subsequent processing and accuracy assessment steps.

The main goal of the QC/QA procedures was the verification of the completeness, logical consistency,

geometric and thematic accuracy and that metadata are following ISO standards on geographic data

quality and INSPIRE data specifications. These assessments were recorded in Data Quality Sheets

which are provided in Annex 2. The QC/QA procedures were based on an assessment of a series of

relevant data elements and processing steps which are part of the categories listed below:

Product requirements;

Specifications of input data: EO data, in-situ data, ancillary data;

Data quality checks: EO data quality, in-situ data quality, ancillary data quality;

Geometric correction, geometric accuracy, data fusion (if applicable), data processing;

Thematic processing: classification, plausibility checks;

Accuracy: thematic accuracy, error matrices

Delivery checks: completeness, compliancy with requirements

After each intermediate processing step a QC/QA was performed to evaluate products appropriateness

for the subsequent processing (see Figure 7).

Figure 7: Quality Control process for EO4SD-Urban product generation. At each intermediate processing

step output properties are compared against pre-defined requirements.




After the initial definition of the product specifications (output) necessary input data were defined and

acquired. Input data include all satellite data and reference data e.g. in-situ data, reference maps,

topographic data, relevant studies, existing standards and specifications, statistics. These input data were

the baseline for the subsequent processing and therefore all input data had to be checked for

completeness, accuracy and consistency. The evaluation of the quality of input data provides

confidence of their suitability for further use (e.g. comparison with actual data) in the subsequent

processing line. Data processing towards the end-product required multiple intermediate processing

steps. To guarantee a traceable and quality assured map production the QC/QA assessment was

performed and documented by personnel responsible for the Quality Control/Assurance. The results of

all relevant steps provided information of the acceptance status of a dataset/product.

The documentation is furthermore important to provide a comprehensive and transparent summary of

each production step and the changes made to the input data. With this information the user will be able

to evaluate the provided services and products. Especially the accuracy assessment of map products and

the related error matrices are highly important to rate the quality and compare map products from

different service providers.

The finalised QC/QA forms are attached in Annex 2.

3.8 Metadata

Metadata provides additional information about the delivered products to enable it to be better

understood. In the current project a harmonised approach to provide metadata in a standardised format

applicable to all products and end-users was adopted. Metadata are provided as XML files, compliant

to the ISO standard 19115 "Metadata" and ISO 19139 "XML Scheme Implementation". The metadata

files have been created and validated by the GIS/IP-operator for each map product with the Infrastructure

for Spatial Information in Europe (INSPIRE) Metadata Editor available at: http://inspire-

geoportal.ec.europa.eu/editor/.

The European Community enacted a Directive in 2007 for the creation of a common geo-data

infrastructure to provide a consistent metadata scheme for geospatial services and products that could

be used not only in Europe but globally. The geospatial infrastructure called INSPIRE was built in a

close relation to existing International Organization for Standardization (ISO) standards. These are ISO

191115, ISO 19119 and ISO 15836. The primary incentive of INSPIRE is to facilitate the use and

sharing of spatial information by providing key elements and guidelines for the creation of metadata for

geospatial products and services.

The INSPIRE Metadata provides a core set of metadata elements which are part of all the delivered geo-

spatial products to the users. Furthermore, the metadata elements provide elements that are necessary to

perform queries, store and relocate data in an efficient manner. The minimum required information is

specified in the Commission Regulation (EC) No 1205/2008 of 3 December 2008 and contains 10

elements:

Information on overall Product in terms of: Point of contact for product generation, date of

creation

Identification of Product: Resource title, Abstract (a short description of product) and Locator

Classification of Spatial Data

Keywords (that define the product)

Geographic information: Area Coverage of the Product

Temporal Reference: Temporal extent; date of publication; date of last revision; date of

creation

Quality and Validity: Lineage, spatial resolution

Conformity: degree of conformance to specifications

Data access constraints or Limitations

Responsible party: contact details and role of contact group/person

These elements (not exhaustive) constitute the core information that has to be provided to meet the

minimum requirements for Metadata compliancy. Each element and its sub-categories or elements have

http://inspire-geoportal.ec.europa.eu/editor/

http://inspire-geoportal.ec.europa.eu/editor/




specific definitions; for example in the element “Quality” there is a component called “Lineage” which has a specific definition as follows: “a statement on process history and/or overall quality of the spatial data set. Where appropriate it may include a statement whether the data set has been validated or quality

assured, whether it is the official version (if multiple versions exist), and whether it has legal validity.

The value domain of this element is free text,” (INSPIRE Metadata Technical Guidelines, 2013). The detailed information on the Metadata elements and their definitions can be found in the “INSPIRE Metadata Implementing Rules: Technical Guidelines,” (2013). Each of the EO4SD-Urban products will

be accompanied by such a descriptive metadata file. It should be noted that the internal use of metadata

in these institutions might not be established at an operational level, but the file format (*.xml) and the

web accessibility of data viewers enable for the full utility of the metadata.




4 Analysis of Mapping Results

This Chapter will present and assess all results which have been produced within the framework of the

current project, in the context of presentation of the Urban Extent product, the LU/LC products and the

Transport Infrastructure product. Furthermore the Sections that follow will provide the results of some

standard analytics undertaken with these products including the following:

Urban Extent – Developments from 2000, 2005, 2010 to 2015

Land Cover Land Use - Status and Trends between 2005 and 2015

Transport Infrastructure - Status and Change between 2005 and 2015

Urban Green Areas – Status and Change between 2005 and 2015

Planned and Unplanned Settlement Areas – Status and Change between 2005 and 2015

Population Distribution and Density – Status and Change between 2005 and 2015

It is envisaged that these analytics provide information on general trends and developments in the Core

and Peri-Urban areas which can then be further interpreted and used by Urban planners and the City

Authorities for city planning.

It should be noted that all digital data sets for these products are provided in concurrence with this City

Report with all the related metadata and Quality Control documentation

4.1 Urban Extent – Developments 2000, 2005, 2010 and 2015

The Urban Extent product in the EO4SD-Urban project is provided by the German Aerospace Centre

(DLR) and is provided for 4 points in time; the 2015 Global Urban Footprint (GUF) Plus product has

been produced jointly exploiting multi-temporal 30m Landsat-8 and ESA Sentinel-1 data with 10m

resolution acquired in 2014-2015. And for the years 2000, 2005 and 2010, the Urban Extent products

generated – given the unavailability of freely and easily accessible multi-temporal radar data at high

resolution – were based only on multi-temporal 30m Landsat-5 and Landsat-7 imagery, and scaled up

to 10m resolution.

As the Urban Extent 2015 product was based on the ESA Sentinel-1 dataset which is a Synthetic

Aperture Radar (SAR) in C band it should be noted that some structures which are flat in nature such as

airport runways were not classified; this is due to the fact that radar relies on backscatter which is more

prominent from vertical features. The GUF+ 2015 products will be validated and available as public

domain data from October 2017 onwards on the Urban Thematic Exploitation Platform (TEP) supported

by the DLR.

In the current project the Urban Extent product for Arusha was first used to assess historical

developments from 2000-2015. Further analysis by overlaying administrative boundaries can be

performed to assess urbanisation extent patterns based on administrative units.

Results:

The first result provided using the different Urban Extent products from 2000 to 2015 is illustrated in

Figure 6 which shows the Urban development in the Core and Peri-Urban areas as well as surrounding

regions of Arusha. The Urban Extent developments after 2000 can be examined by Urban Planners to

identify different patterns of growth such as “Edge Growth” or “Leapfrog Growth” depending on the

location of the developments.

From the depiction of urban extent developments between 2000 and 2015 in Figure 8 and Figure 9, one

can note that urban extent development occurred in Arusha in large blocks around existing residential

areas and also stretches in two parallel lines towards south. The urban development till 2005 stretches

towards east in very small fragmented patches. The same patterns appear in the next epoch 2005 to 2010

whereas additionally larger built-up areas occur in the western as well as in the eastern outskirts of

Arusha. In 2010, the urban extension further increased at the surrounding, but most of the extension

occurred in the very south of the core urban area and stretches in very small fragmented patters far to

the east.




Figure 8: Urban Extent developments in the epochs 2000 to 2005, 2005 to 2010 and 2010 to 2015 in Arusha

and surrounding region.

Figure 9: Urban Extent developments in the epochs 2000 to 2005, 2005 to 2010 and 2010 to 2015 in Arusha

within the Core Urban Area.




4.2 Land Cover Land Use 2005 and 2015

This Section will present the results of the LU/LC mapping for 2005 and 2015 as well the statistical

information on the changes between these two epochs. The LU/LC overview map for 2015 is depicted

in Figure 10 and a cartographic version of the map layout is provided as a pdf file in addition to the geo-

spatial product.

Figure 10: Left: Overview of the Service Area. Right: Detailed Land Cover Land Use 2015 in Arusha.

In the 2005 epoch the most dominant LU/LC classes occurring in the Overall area were: Agriculture

(61.28% of the total area), Residential (20.0 % of the total area) and Forest areas (9.06% of the total

area). By 2015 these classes remained as the most dominant with some slight changes. For instance, the

Residential and the Forest area classes increased by 3.73% and 0.38% (2.4 km2) respectively. Further

information on the class disaggregation and area coverage is presented in Figure 11 and Figure 12 for

the epochs 2005 and 2015 respectively.




Figure 11: Detailed Land Cover Land Use 2005 structure: Presented as Overall, Core Urban and in Peri-

Urban in % (left) and km2 (right).

Similarly to the Overall area, Agriculture was the most dominant class in 2005 in both the Core and the

Peri-Urban areas with 56.48% and 63.72%, respectively. The main difference in these two areas is the

coverage and density of the Residential class, and the extent of the commercial and industrial areas. The

Residential areas in the Peri-Urban belong to the low density classes and are scattered over larger areas.

By 2015 (see Figure 12), several trends were observed in the class distribution and the area coverage.

For instance, the Agriculture class decreased by 3.79% for the period 2005 - 2015, by losing 23.68 km2

from its initial area of 382.90 km2 in 2005. The Residential class which represented 20% (124.89 km2)

from the Overall area in 2005, increased by 3.73% thus accounting for 148.21 km2 from the Overall area

by 2015. Detailed information on the area, percentage distribution and changes can be further observed

in Table 12.

Figure 12: Detailed Land Cover Land Use 2015 structure: Presented as Overall, Core Urban and in Peri-

Urban in % (left) and km2 (right).

The next Section will highlight the LU/LC change information between the two epochs in more detail.

Description of LULC Changes:

In addition to the overall LU/LC classification for the two epochs it is interesting to assess the different

trends between classes over the 10 year time period. The quantitative figures for each class (combined

Core and Peri-Urban) are first provided in Table 12 to get an overview.




Table 12: Detailed information on area and percentage of total area for each class for 2005 and 2015 as

well as the changes.

LU/LC Classes 2015 2005 Change Change per Year

sqkm % of

total sqkm

% of

total sqkm % sqkm %

Residential 0-10 % 73.30 11.73% 83.63 13.38% -10.33 -1.65% -1.03 -0.17%

Residential 10-30 % 33.44 5.35% 23.34 3.74% 10.09 1.62% 1.01 0.16%

Residential 30-50 % 22.93 3.67% 9.14 1.46% 13.79 2.21% 1.38 0.22%

Residential 50-80 % 14.86 2.38% 7.38 1.18% 7.48 1.20% 0.75 0.12%

Residential 80-100 % 3.68 0.59% 1.40 0.22% 2.28 0.37% 0.23 0.04%

Industrial, Commercial, Public, Military 11.45 1.83% 9.55 1.53% 1.91 0.30% 0.19 0.03%

Arterial Line 1.37 0.22% 1.37 0.22% 0.00 0.00% 0.00 0.00%

Toll Line 4.46 0.71% 4.46 0.71% 0.00 0.00% 0.00 0.00%

Airport 0.50 0.08% 0.50 0.08% 0.00 0.00% 0.00 0.00%

Mining_QuarryAreas_DumpSites 0.89 0.14% 0.55 0.09% 0.34 0.05% 0.03 0.01%

Construction Site 3.38 0.54% 1.93 0.31% 1.45 0.23% 0.14 0.02%

Land without current use 4.16 0.67% 2.93 0.47% 1.23 0.20% 0.12 0.02%

Urban Parks 0.19 0.03% 0.18 0.03% 0.01 0.00% 0.00 0.00%

Recreation Facilities 0.90 0.14% 0.75 0.12% 0.15 0.02% 0.01 0.00%

Cemeteries 0.06 0.01% 0.04 0.01% 0.02 0.00% 0.00 0.00%

Agricultural Area 359.21 57.49% 382.90 61.28% -23.68 -3.79% -2.37 -0.38%

Forest 59.01 9.44% 56.61 9.06% 2.40 0.38% 0.24 0.04%

Natural areas (non-forested) 29.22 4.68% 36.33 5.82% -7.11 -1.14% -0.71 -0.11%

Bare Soil 0.00 0.00% 0.06 0.01% -0.06 -0.01% -0.01 0.00%

Wetlands 0.01 0.00% 0.01 0.00% 0.00 0.00% 0.00 0.00%

Water 1.78 0.28% 1.74 0.28% 0.04 0.01% 0.00 0.00%

Total 624.81 100.00% 624.81 100.00%

The area statistics of the LU/LC classes show a lot of change dynamics within the Residential density

classes. For example, it can be noted that there is an evident loss of area in the Very Low Density class

(0-10 %) with 1.65% and increase in all other density classes.

4.2.1 Spatial Distribution of Main LU/LC Change Categories

In order to better analyse the growth trend and the spatial distribution of changes meaningful

aggregations of the LU/LC classes in both epochs were used. The following categories were developed:

Urban Densification: Changes from lower Residential Density Class into a higher Residential

Density class;

Urban Residential Extension: all changes from Non-Urban Residential classes to a Residential

class;

Other Urban Land Use Extension: all changes from Non-Residential Urban classes to Other

Urban and Non-Urban classes.

Changes within Natural and Semi-Natural Areas: all changes in between the natural and semi-

natural classes (e.g. Forest into agriculture).

The overlay analysis of these aggregated categories of the epochs 2005 and 2015 is depicted in Figure

13.




Figure 13: Land Cover Land Use Change Types - Spatial Distribution.

The spatial distribution of the change types as depicted in Figure 13 shows that the urban densification

in the 10 year period mainly happened in the city centre of Arusha and in the eastern Peri-Urban area

along the main roads. The Extension of the Residential areas (red colour in Figure 13) occurred mainly

in the south-western part of the Core Urban area. Small scattered areas in the eastern part of the Core

city experienced Residential extension as well.

The statistics of the Change categories are presented in Figure 14 and Table 13. The statistics in Figure

14 provide a quantitative aspects to the LU change classes; for example it’s interesting to note that Urban Densification was the most dominant class, covering 76.45% of the Peri-Urban area, whereas the

Residential Extension accounted for the most changes in the Core Urban area with 45% of all the

changes.

Figure 14: Land Cover Land Use Change Types 2000-2015 - overall, in Core Urban and in Peri-Urban

Zone in % (left) and km² (right) in Arusha.




Regarding the Overall area of Arusha, the most dynamic changes which occurred between 2005 – 2015

were represented by the Urban Densification class, i.e. changes from lower to higher Residential Density

classes with 48.34 % from the total area as presented in Figure 14. The second largest block of changes

can be observed within the Urban Residential Extension, followed by Other Urban Land Use Extension.

A main difference between the Core Urban and the Peri-Urban areas in Arusha was that only 0.27%

from the Peri-Urban Zone was used for residential extension, whereas the land conversion in the Core

Urban Zone was mainly related to Urban Residential Extension with 45%. The quantitative data is

presented in Table 13.

Table 13: Overall LU/LC Statistics.

Change Classes Change Overall Change Core Urban Change Peri-Urban

sqkm % sqkm % sqkm %

Urban Densification 32.28 48.34% 17.32 36.69% 14.96 76.45%

Urban Residential Extension 21.38 32.02% 21.33 45.19% 0.05 0.27%

Other Urban Land Use

Extension 7.58 11.35% 5.06 10.71% 2.53 12.91%

Change within Natural and

Semi-Natural Areas 5.53 8.28% 3.50 7.41% 2.03 10.38%

Total 66.77 100.00% 47.20 100.00% 19.57 100.00%

4.2.2 Changes of Agricultural Areas

In order to analyse the relatively large loss of Agricultural areas as noted in the earlier part of Section

4.2, further change analysis was performed. The following change categories were developed for this

analysis:

Agriculture to Residential area

Agriculture to Industry, Commercial; Public or Military area

Agriculture to Plantations

The spatial distribution of these changes are displayed in Figure 15.

Figure 15: Spatial distribution of changes from Agricultural Areas to other Classes between 2005 and 2015.




The conversion of agricultural areas into other LULC classes was mainly caused by the residential

extension and reforestation in Arusha. These two conversion types are shown in Figure 15 in blue and

green colours respectively. Especially in the south-western area of Arusha, large agricultural areas were

replaced by urban settlements.

Figure 15 and Figure 16 provide a quantitative overview of the changes of the agricultural areas into

other LU classes between 2005 and 2015 in Arusha. As it can be observed in Figure 15, most of the

agricultural areas subject to change were converted to residential areas (blue colour). A few larger areas

in the western part of Arusha were converted into industrial or commercial land use (red colour). Within

the core region most of this conversion type are scattered through-out the entire city (see Figure 15, red

colour). The third change class in Figure 15 depicts changes from agricultural to forested areas, which

only have a minor contribution to the LULC changes in the period between 2005 - 2015, but they are

the main change type in the Peri-Urban area. The percentage of each change type is presented in Figure

16.

Figure 16: Changes of Agricultural Areas into other LU classes between 2005 and 2015; Presented as

Overall, Core Urban and in Peri-Urban Zone in % (left) and km2 (right).

The statistics in Figure 16 and Table 14 show that the majority of Agricultural areas in the Core Urban

has been converted into Residential area in the 2005 – 2015 period. In the Peri-Urban Zone the

agricultural areas mainly changed to residential areas, followed by forested land.

Table 14: Statistics of changes of Agricultural areas.

Change Classes Change Overall Change Core Urban Change Peri-Urban

sqkm % sqkm % sqkm %

Agricultural Area to

Residential 19.68 88.24% 19.64 88.41% 0.04 46.12%

Agricultural Area to Industrial,

Commercial, Public or Military

Area

1.39 6.24% 1.38 6.23% 0.01 8.25%

Agricultural Area to Forest 1.23 5.52% 1.19 5.36% 0.04 45.62%

Total 22.30 100.00% 22.21 100.00% 0.09 100.00%




4.3 Transport Network

The Transport Network was created for both points in time (2005 and 2015) using three road types. The

Arterial roads and Collector roads were integrated in the LULC map by applying a buffer of 12 m and

8 m for the Arterial and Collector roads, respectively. Local roads are only part of the vector data set,

which are provided to the user.

Figure 17 depicts the Transport Network for both points in time. The left figure presents the Transport

Network in 2005 and the right the Transport Network in 2015. The main changes of the Transport

Network occurred in terms of densification of Local roads in 2015 in the north-western part and in the

south-western part of the Core Urban area.

Figure 17: Transport Network of Arusha in 2005 and 2015.

4.4 Urban Green Areas

The location and extent of green areas are determined within the product of urban land use/ land cover

at Level I. Urban green areas refer to land within and on the edges of a city that is partly or completely

covered with grass, trees, shrubs, or other vegetation. The product delivered provides accurate

information (1 m resolution) on the spatial location and extent of green areas located within the Urban

Extent (Level I class: 1000) derived from the baseline LULC information product. Detecting and

monitoring urban green coverage needs very high resolution optical satellite images, which explains the

product generation over the Core Urban Area of Arusha only.

The overview map in Figure 18 gives an overview of the mapping result and the structure of the map,

which is designed for a print out at DIN A0 paper size (84.1 x 118.9 cm).




Figure 18: Map overview of Urban Green Areas in Arusha. Green area loss, gain and stable green areas

can be identified. The map is delivered as separate product of high resolution for printing at

paper size DIN A0, which is 84.1 cm x 118.9 cm.

The spatial distribution of the change types as depicted in Figure 18 shows that changes are distributed

across the entire city centre area of Arusha. The surrounding areas are mostly covered by Agriculture

and/or Natural or semi-Natural areas. Overall 24.79% of the entire area was covered by vegetation in

2006 and 2016. The percentage of gain of green area (15.66%) is a little bit higher than the percentage

of loss of green area (14.34%) resulting in an increase of urban green areas in Arusha. About 45.20% of

the overall area was covered by artificial class (Level I class).

Figure 19: Percentage of urban green areas within the core area of Arusha. The pie chart illustrates the

status and change of urban green areas in-between 2005 and 2015.




The changes within the urban green areas can be illustrated as an overall area coverage. Figure 20

represents the green area coverage in square kilometers in 2005 and 2015, compared to the amount of

the artificial areas class. Both classes demonstrated a relatively equal growth within the 10 year period.

Urban areas showed an increase of 17.65 km² (from 37.7 km² to 55.35 km²) and green areas an increase

of 8.27 km² (from 24.17 km² to 32.44 km²).

Figure 20: Bar charts for both points in time presenting the total area of urban greenery versus non-

green areas.

4.5 Planned and Unplanned Settlement Areas

The location and extent of planned and unplanned settlement areas are determined within the classes

“1100 Residential” and “1211 Commercial Areas” of the urban land use/ land cover product. Polygons

with the LU class “Commercial Areas” were only included for classification in this product, when they also contained residential buildings.

In an effort to use spatial patterns to depict the class of “planned settlements” (see Figure 21) there was

a focus on residential areas where the houses are oriented in the same direction, similar in size and shape

and have a well-defined road network, often in a grid pattern which provides almost direct access to

every house. For “unplanned settlements” (see Figure 22) houses are not oriented in the same direction

and streets, if existing are curved, small and appear to have been constructed in an ad hoc manner (no

distinct patterns). Also not every building can be reached by a road.

To further support the classification effort, the Arusha Master Plan 2015-2035, Volume 2: Technical

Supplement 1-6 (2016) page 82 was used to help in orienting the locations of planned and unplanned

settlement areas.

0

10

20

30

40

50

60

2005 2015

are

a i

n k

m²

urban (km2) green (km2)




Figure 21: Planned settlement areas in Arusha in 2015.

Figure 22: Unplanned settlement areas in Arusha in 2015.




Due to the complexity of identifying these classes, the detection and monitoring planned and unplanned

settlements needs VHR optical satellite images; thus the product was generated only over the Core

Urban Area of Arusha.

The overview map in Figure 23 shows the changes in planned and unplanned settlement areas in Arusha

between 2005 and 2015. The map is delivered as separate product of high resolution for printing at paper

size DIN A0, which is 84.1 cm x 118.9 cm. gives an overview of the mapping result and the structure

of the map, which is designed for a print out at DIN A0 paper size (84.1 x 118.9 cm).

Figure 23: Map overview of changes in planned and unplanned settlement areas in Arusha during the years

2005 and 2015.

Figure 24 shows the spatial distribution of planned and unplanned areas between the two years 2005

and 2015 in Arusha. Most of the city is covered by unplanned settlement areas, only the inner city as

well as a part stretching out to the south has planned settlements. The overview map also shows that

most of the new developed settlements are unplanned.

An analysis of the changes is depicted in Figure 24. The pie chart shows that more than half of the Core

Urban area has unplanned settlements (52.17%), and only 11.53% of the residential area is planned.

During the 10 years the unplanned settlement area further grew by 31.53%, while the planned settlement

area only grew by 3.96%.




Figure 24: Percentage of planned and unplanned areas within the core area of Arusha. The pie chart

illustrates the status and change of planned and unplanned areas in-between 2005 and 2015.

The changes in the overall area coverage for both the planned and unplanned settlement areas are

illustrated in Figure 25; the bar chart represents the planned and unplanned settlement areas in square

kilometres in 2005 and 2015. Both unplanned and planned settlement increased in area within the 10

years, however the unplanned settlement area which was already more than half of the whole residential

area increased in area more than the planned settlement area. In 2005 the planned settlement area

covered 8.14km2, while the unplanned settlement area covered 36.59km2. In 2015, the planned

settlement area increased to 11.20km2, and the unplanned settlement to 57.52km2.

Figure 25: Bar charts for both points in time presenting the total area of planned and unplanned settlement

areas.

52.17%

11.53%

31.53%

3.96% 0.28% 0.32% 0.81%

No Change in Unplanned Settlement Area No Change in Planned Settlement Area

Expansion of Unplanned Settlement Area Expansion of Planned Settlement Area

Decrease of Unplanned Settlement Area Decrease of Planned Settlement Area

Unplanned to Planned Settlement Area

0

10

20

30

40

50

60

70

2005 2015

Are

a [

km2

]

Unplanned Settlement Area Planned Settlement Area




4.6 Population Distribution and Density

The Population Distribution and Density Change product contains spatial explicit information about

population distribution within the Core Urban Districts of Arusha. The product is derived for all units

which belong to the residential class (LULC class 11), using five different data sources: 1) Land

Use/Land Cover Baseline Product; 2) WorldPop data (2015) with a spatial resolution of 100m; 3)

Official Population census data from 2002 and 2012; 4) Imperviousness/Soil Sealing layer (see Annex

1 for a full description of this product) and 5) Administrative boundaries. An important note is that the

product has been always developed for the period 2005 – 2015, due to data availability (the global

WorldPop dataset was only available only for the year 2015), and to meet the user requirements,

regarding time interval of 10 years between the historic and current state of the product.

The overview map in Figure 26 shows eight different change classes which were identified, based on a

frequency distribution analysis of the changes that occur: Unchanged Population Distribution; Up to –100% decrease; Up to 200% increase; 201% - 400% increase; 401% - 600% increase; 601% - 800%

increase; 801% - 1000% increase; More than 1000% increase. The Unchanged Population Distribution

class is defined as the residential units which experienced stable population distribution between 2005

and 2015, with values within the overall annual population growth rate (-4% to 4%). The map is

delivered as separate product of high resolution for printing at paper size DIN A0, which is 84.1 cm x

118.9 cm. It has to be noted that the GIS data set allows map printouts at any other scale which however

would require different map sheet layouts and cartographic designs.

Figure 26: Overview Map of Population Distribution Change in Arusha (2005 – 2015).

The spatial distribution of the change types as depicted in Figure 26 can be grouped in 3 main categories:

1) unchanged areas; 2) areas which experienced population decrease and 3) areas which increased in

terms of population distribution at a different degree. The overview map shows that most of the

residential units in Arusha experienced increase of up to 200% (depicted in light yellow colour). The

decrease in population is heterogeneously distributed across all Core Urban Districts, whereas the

highest degree of increase (More than 1000%) is observed mainly on the east, which is closely linked

to urban expansion and densification in this area.




An analysis of the change types depicted in Figure 27, shows that 64.09% of the area was subject to

population increase of up to 200%, followed by decreasing population which accounted for 13.41% of

the total area. Only 1.5% remained unchanged (within the range of the annual growth rate of -4% to

4%).

Figure 27: Population Distribution Change within the Core Urban Districts of Arusha between 2005 and

2015.

Figure 28 provides further detail on the population distribution changes between 2005 and 2015, in

relation to build up areas which are based on the amount of soil sealing degree in each Core Urban

District.

Figure 28: Changes in Population Distribution, in relation to soil sealing degree in Arusha between 2005

and 2015.




For instance, it is noted that in Sokoni I (the biggest district in terms of population) more than 25000

people moved to build up areas with a sealing degree of 50 – 80%. Similarly, a high proportion of the

residents in Elerai, Sombetini and Unga Ltd moved to areas with sealing degree of 50 – 80%. The

increase of inhabitants in most of the Core Urban Districts was proportionally distributed among the

different soil sealing classes. In contrast, the Very High Density class (80 – 100%) experienced decrease

in most of the districts. This trend can be driven by different factors, such as: rent price in central

districts, lack of land availability, and conversion of residential units to commercial, industrial, central

business districts in the city centre, etc. However, the information presented in Figure 28 cannot

explicitly define those driving forces. Therefore, further research is needed to identify and justify the

drivers behind such phenomena, thus helping urban planners and local officials in the decision – making

process.

4.7 Concluding Points

This Chapter 4 presented only a summary and overview of what is possible in term of analytics with the

geo-spatial datasets provided for Arusha in the current project. This Report is a living document and

will be complemented with further analysis during the project. Important would be to further analyse

the EO4SD Urban datasets with the City Master Plan for Arusha in order to enhance the latter for

planning purposes.




5 References

Arusha Master Plan 2015-2035 (2016). Ministry of Lands, Housing and Human Development,

Tanzania.

Czaplewski, R. L. (2003). Chapter 5: Accuracy assessment of maps of forest condition: statistical design

and methodological considerations, pp. 115–140. In Michael A. Wulder, & Steven E. Franklin (Eds.),

Remote sensing of forest environments: concepts and case studies. Boston: Kluwer Academic Publishers

(515 pp.).

European Union (2011). Mapping Guide for a European Urban Atlas, Version 11.0

Goodchild, M., Chih-Chang, L. and Leung, Y. (1994): Visualizing fuzzy maps, pp. 158-67. In

Heamshaw, H.H. and Unwin, D.J. (Eds.), Visualization in geographical information systems.

Chichester: Wiley.

Olofsson, P., Foody, G. M., Stehman, S. V., & Woodcock, C. E. (2013). Making better use of accuracy

data in land change studies: Estimating accuracy and area and quantifying uncertainty using stratified

estimation. Remote Sensing of Environment, 129, 122–131. doi:10.1016/j.rse.2012.10.031

Selkowitz, D. J., & Stehman, S. V. (2011). Thematic accuracy of the National Land Cover Database

(NLCD) 2001 land cover for Alaska. Remote Sensing of Environment, 115(6), 1401–1407.

doi:10.1016/j.rse.2011.01.020.

Internet

Road classification, European Commission, 2017, https://ec.europa.eu/transport/road_safety/specialist-

/knowledge/road/designing_for_road_function/road_classification_en, last accessed 2017.08.17

Earth Observation for Sustainable

Development Urban Project QA/QC Sheets developed by GAF AG

© 2018 GAF AG All Rights Reserved. Unless otherwise indicated, the templates of these QA/QC pages are copyrighted by GAF AG. No part of these pages, either text or image may be used for any purpose other than use in the EO4SD-Urban Project. Therefore, reproduction, modification, storage in a retrieval system or retransmission, in any form or by any means, electronic, mechanical or otherwise, for reasons other than personal use, is strictly prohibited without prior written permission

Page 1





Page 2

Annex 1 – Processing Methods for EO4SD-Urban Products




Page 3





Page 4

Summary of Processing Methods

Urban and Peri-Urban Land Use/Land Cover and Change

The input includes Very High Spatial Resolution (VHR) imagery from different sensors acquired at

different time. The data is pre-processed to ensure a high level of geometric and radiometric quality

(ortho-rectification, radiometric calibration, pan-sharpening).

The complexity when dealing with VHR images comes from the internal variability of the information

for a single land-use. For instance, an urban area is represented by a high number of heterogeneous pixel

values hampering the use of automated pixel-based classification techniques.

For these VHR images, it is possible to identify textures (or pattern) inside an entity such as an

agricultural parcel or an urban lot. In other words, whereas pixel-based techniques focus on the local

information of each single pixel (including intensity / DN value), texture analysis provides global

information in a group of neighbouring pixels (including distribution of a group intensity / DN values

but also spatial arrangement of these values). Texture and spectral information are combined with a

segmentation algorithm in an Object Based Image Analysis (OBIA) approach to reach a high degree of

automation for most of the peri-urban rural classes. However, within urban land, land use information

is often difficult to obtain from the imagery alone and ancillary/in situ data needs to be used. The

heterogeneity and format of these data mean that another information extraction method based on

Computer Aided Photo-Interpretation techniques (CAPI) need to be used to fully characterise the LULC

classes in urban areas. Therefore, a mix of automated (OBIA) and CAPI are used to optimise the

cost/quality ratio for the production of the LULC/LUCC product. The output format is typically in vector

form which makes it easier for integration in a GIS and for subsequent analysis.

Level 4 of the nomenclature can be obtained based on additional information. These can be generated

by more detailed CAPI (e.g. identification of waste sites) or by an automated approach based on

derived/additional products. An example is illustration by categorising the density of the urban fabric

which is related to population density and can then subsequently used for disaggregating population

data.

Information on urban fabric density can be obtained through several manners with increasing level of

complexity. The Imperviousness Degree (IMD) or Soil Sealing (SL) layer (see separate product) can be

produced relatively easily based on the urban extent derived from the LULC product and a linear model

between imperviousness areas and vegetation vigour that can be obtained from Sentinel 2 or equivalent

NDVI time series. This additional layer can be used to identify continuous and discontinuous urban

fabric classes. Five urban fabric classes can be extracted based on a fully automated procedure:

Continuous urban fabric (IMD > 80%)

Discontinuous dense urban fabric: (IMD 50-80 %)

Discontinuous medium density urban fabric (IMD: 30-50 %)

Discontinuous low density urban fabric (IMD 10-30 %)

Discontinuous very low density urban fabric (IMD < 10 %)

Manual enhancement is the final post-processing step of the production framework. It will aim to

validate the detected classes and adjust classes’ polygon geometry if necessary to ensure that the correct

MMU is applied. Finally, a thorough completeness and logical consistency check is applied to ensure

the topological integrity and coherence of the product.




Page 5

Change detection: Four important aspects have to be considered to monitor land use/land cover change

effectively with remote sensing images: (1) detecting that changes have occurred, (2) identifying the

nature of the change, (3) characterising the areal extent of the change and (4) assessing the spatial pattern

of the change.

The change detection layer can be derived based on an image-to-image approach provided the same

sensor is used. An original and efficient image processing chain is promoted to compare two dates’ images and provide multi-labelled changes. The approach mainly relies on texture analysis, which has

the benefits to deal easily with heterogeneous data and VHR images. The applied change mapping

approach is based on spectral information of both dates’ images and more accurate than a map-to-map

comparison.




Page 6


Urban Extent and Change

Reliably outlining urban areas is of high importance since an accurate characterization of the urban

extent is fundamental for accurately estimating, among others, the population distribution, the use of

resources (e.g., soil, energy, water, materials), infrastructure and transport needs, socioeconomic

development, human health and food security. Moreover, monitoring the change in the extent of urban

areas over time is of great support for properly modelling the spatial-temporal patterns of urbanisation

evolution and, thus, better estimating future trends and implementing suitable planning strategies.

The product is a binary mask outlining in the area of interest the urban areas (intended as built-up

structures) with respect to all other land-cover classes merged together into a single information class.

The urban class and the non-urban class are associated with value “255” and “0”, respectively. Regrouping of relevant LULC thematic classes can be used to depict urban extent precisely. Instead, if

a detailed LULC product is not available for the selected study region, then the information will be

derived by the following approach.

The product is generated at 30 m spatial resolution by properly exploiting Landsat-4/5/7/8 multi-

temporal imagery acquired over the peri-urban and urban area within a given time interval of interest in

which no relevant changes are expected to occur (typically a time period of 1-2 years allows to obtain

very accurate results). For all the considered scenes, cloud masking and, optionally, atmospheric

correction are performed. Next, a series of features specifically suitable for delineating urban areas are

derived for each image. These include both spectral indexes (e.g., the normalized different vegetation

index (NDVI), the atmospherically resistant vegetation index (ARVI), the normalized difference water

index (NDWI), etc.) and texture features (e.g., occurrence textures, co-occurrence texture, local

coefficient of variation, etc.). The core idea is then to compute per each pixel key temporal statistics for

all the extracted features, like temporal maximum, minimum, mean, variance, median, etc. This allows

compressing all the information contained in the different multi-temporal acquisitions, but at the same

time to easily and effectively characterize the underlying dynamics. It is worth noting that for different

pixels in the study area, different number of scenes might be available. However, in the hypothesis of a

sufficient minimum number of acquisitions for computing consistent statistics, this does not represent

an issue. Moreover, in this framework it is also possible to obtain spatially consistent datasets to be

employed for the desired analyses even when investigating large areas. Training data for the urban and

non-urban class are then extracted by employing a strategy based on the analysis of the DLR Global

Urban Footprint (GUF) layer (which varies depending whether the target period of interest refers to a

time interval before or after that which the GUF refers to). Afterwards, a Support Vector Machines

(SVM) classifier is employed where a Radial Basis Function (RBF) kernel is used.




Page 7


Transport Network

The transportation network is mainly manually digitised and mapped on the basis of very high resolution

(VHR) optical satellite imagery. Requested features are obtained by integration of auxiliary data such

as OpenStreetMap (OSM) or local datasets as a starting point for the product generation. The revision

and update of the auxiliary data is realised by using up-to-date VHR data. The workflow can be specified

as follows:

Quality check of available ancillary data such as local data sets.

Processing of optical satellite data – dependent on satellite data product level (geometric,

atmospheric and radiometric corrections, enhancements – colour optimization, mosaicking,

tiling).

Identification, collection and integration of available ancillary data (e.g. Open Street Map)

Identification and adjustment of spatial inconsistencies. The OSM data is used as spatial

reference. Upon User request other data sets can be used.

Update of the network by visual photo-interpretation according road hierarchy (see description

below).

Update of attributes by photointerpretation.

Generalization, application of MMU (minimum allowable dangling length)

Quality control and accuracy assessment

o Statistical sampling of check points

o Independent evaluation of products (second interpreter, third party assessment)

Change detection

The road hierarchy used in the classification is based on the international road classification

standards. One definition is specified by the European Commission. Roads are divided into three

groups - arterial or through traffic flow routes (in our case Arterial Roads), distributor road (in our

case Collector Roads), and access roads (in our case Local Roads). The three road types are defined

as follows: Arterial Roads - roads with a flow function allow efficient throughput of (long distance)

motorized traffic. All motorways and express roads as well as some urban ring roads have a flow

function. The number of access and exit points is limited. Collector Roads - roads with an area

distributor function allow entering and leaving residential areas, recreational areas, industrial zones,

and rural settlements with scattered destinations. Local Roads - roads with an access function allow

actual access to properties alongside a road or street. Arterial roads and collector roads were the

main focus of the classification. These types of roads were identified for the entire Area of Interest.




Page 8


Urban Green Areas

The location and extent of green areas are determined within the product of urban land use/ land cover

at Level I. Urban green areas refer to land within and on the edges of a city that is partly or completely

covered with grass, trees, shrubs, or other vegetation. This includes public parks, private gardens,

cemeteries, forested areas as well as trees, river alignments, hedges etc. The product delivered within

EO4SD-Urban project thus provides accurate information (1 m resolution) on the spatial location and

extent of the green areas located within the Urban Extent (Level I class: 1000) derived from the baseline

LULC information product.

Detecting and monitoring urban green coverage needs very high resolution optical satellite images,

which explains the product generation over the Core Urban Area of AOI only. The same images have

been logically used for generating the LULC information product. Consequently, the usual preliminary

quality check and pre-processing tasks were already implemented.

Urban Green Areas have been detected using automated non-supervised classification method. More

precisely, each single multispectral VHR scene has been classified by specifying the most appropriate

algorithm and class number. Then, pixel units from the classes considered as representing green areas

have been combined into 1 single class. From this operation results the required binary raster product.

At this stage, it only remains necessary to apply some post-processing steps:

Morphological filter is applied to fill small gaps within the green areas (caused by shadow)

Resampling of the data to the provided spatial resolution of 1m

Removing small pixel groups under the minimum mapping unit.

Integrating the information provided by the LULC product (e.g. class Urban Parks, Cemeteries).

Validation of Mapping results

Furthermore, using archive very high resolution images, current and historic extent of urban green areas

are compared to identify their temporal evolution – extent growth or reduction. Quality control and

accuracy assessment tasks are performed by means of visual interpretation considering also the LULC

dataset.




Page 9


Planned and Unplanned Settlements

The location and extent of planned and unplanned settlement areas are determined within the classes

“1100 Residential” and “1211 Commercial Areas” of the urban land use/ land cover product. Polygons

with the LU class “Commercial Areas” were only included for classification in this product, when they also contained residential buildings.

Two distinct between the two types following rules were applied:

In planned settlement areas (see Figure 1 left side) houses are oriented in the same direction, are

similar in size and shape and show a good road network reaching every house.

In unplanned settlements (see Figure 1 right side) houses are not oriented in the same direction

and streets, if existing are curved and small. Not every building can be reached by a road.

Figure 1: Planned (left) and unplanned (right) settlement areas in Arusha in 2015.

To further improve the results, the City Urban Master plans were used to help in the distinction in

planned and unplanned settlement areas. Detecting and monitoring planned and unplanned settlements

need very high resolution (VHR) optical satellite images; therefore this product was only generated over

the Core Urban Area of Arusha only. As the same images VHR images were used for generating the

LULC information product, the usual preliminary quality check and pre-processing tasks were already

implemented. Planned and unplanned areas have been detected by visual interpretation of the actual and

historic VHR image by following the rules described above and with the help of the City Master Plans

of the three cities.

Quality control and accuracy assessment tasks are also performed by means of visual interpretation

considering also the LULC dataset.




Page 10



This product aims at providing information on the spatial distribution of residents in a specified Area of

Interest (AOI). Additionally, it depicts: a) location of people; b) estimated population in a specific spatial

unit; and c) changes in population distribution throughout the years.

Data sources and types

To derive the Population Distribution and Density product, five main ancillary data types are used: 1)

Land Use/Land Cover Baseline Product; 2) WorldPop data; 3) Population census data; 4)

Imperviousness/Soil Sealing layer produced by DLR and 5) Administrative boundaries.

1) LULC product is used to extract the residential units within the Core Urban Districts of the city.

2) WorldPop data is a globally available dataset with spatial resolution of 100m, which provides

disaggregated population counts at a specific spatial unit (on a pixel level) for the year 2015.

Therefore, this dataset is used to interpolate the population counts for each residential unit in

the current state of the product, whereas the historic state is modelled, based on soil sealing

information and official population census data.

3) Population census data is further applied to improve the quality of the current state of the product

by adjusting population residuals, and to extrapolate the population size in the historic state of

the product.

4) Imperviousness/Soil Sealing layer has been produced by the German Space Agency (DLR). The

product provides for each pixel identified as urban the corresponding estimated Percentage

Impervious Surface (PIS). When used with the population data this PIS layer supports the

disaggregation of the population counts from district level to pixel level, using the mean sealing

degree for each urban fabric class.

5) Administrative boundaries (district level) are downloaded from the Global Administrative

Areas Website. Furthermore, they are clipped with the specified AOI and used to inform the

total number of inhabitants in each district, and to estimate the overall population density on a

district level.

Methodology

This section describes the general structure of the methodology which is used to compute Population

Distribution and Density for the specific study period. It covers the following two parts: 1) Interpolation

and projection of the population data; and 2) Spatial disaggregation and adjustment of the population

residuals.

1. Interpolation and projection of the population data

The first step from this methodological approach is to interpolate and to project the total number of

inhabitants per district for the specific reference year, based on official population census data. To

support this analysis, population census datasets from 2002 and 2012 are acquired from the Integrated

Public Use Microdata Series (IPUMS) and the United Republic of Tanzania, Ministry of East African

Cooperation, in order to estimate the population size in the Core Urban Districts of Arusha for 2005 and

2015, respectively (IPUMS; MEAC, 2013) . First, the overall annual population growth rate is calculated

based on the total population size, using the following formula as in Kindu et al. (2015): 𝑃 = 𝑃 𝑒 (1)

where: 𝑃0 is the total population in 2002, 𝑃 is the total population in 2012,




Page 11

represents the number of years between the two periods,

is the average annual growth rate.

Next, the average annual population growth rate per district is derived, based on the available census

data (2002 – 2012). Thus, the population size for each district is projected and estimated for 2015, using

equation (1). Finally, the population size in each district in 2005 is interpolated, based on the average

annual population growth rate per district or the overall annual growth rate (when census data for the

initial reference year (2002) was not available).

The population census datasets are used as a supportive parameter in the computation of the current state

of the product, in order to re-distribute the population residuals, created from the interpolation of the

WoldPop data. In contrast, it is the main source for population disaggregation in the historic state, since

no historic WorldPop data is available.

2. Spatial disaggregation and adjustment of the population residuals

In order to improve the quality of the current state of the product (based on WorldPop data), and to

interpolate the historic, further disaggregation analysis is conducted, following the general approach of

Batista e Silva et al. (2013): 𝐾 = 𝑃𝑆∑ 𝑈𝑐∗𝑆𝑐 𝑐 (2)

where: 𝐾 represents the number of inhabitants per unit (%) of sealed surface in each district,

Ps accounts for the total population in the specific district, based on the official census data,

Uc is the number of urban pixels (in raster data), or the sum of class area within the district (in

a vector data),

Sc is the mean soil sealing degree of each urban fabric class c.

Next, the estimated number of residents for each urban fabric class c (𝑃𝑐 is derived, as in Batista e

Silva et al. (2013): 𝑃𝑐 = 𝑘 ∗ 𝑆𝑐 (3)

Finally, a dasymetric population technique is applied to disaggregate the number of residents per spatial

unit and to adjust the population residuals. This approach takes into consideration the residential

polygons from the LULC product, denoted as ‘target’ zone, and the number of inhabitants per district, based on the official census data as ‘source’ zone, using the following formula (Batista e Silva, et al. 2013, p.17): 𝑃 ’ = 𝑃 ∗ 𝐴𝑖∗𝑊𝑖∑𝑖 𝐴𝑖∗𝑊𝑖 (4)

where: 𝑃𝑖′ refers to the estimated population in the target zone i, 𝑃 is the known population in the ‘source’ zone s

Ai is the area of the ‘target’ zone polygons

Wi is the weighting parameter related to the population density of the ‘target’ zone. n corresponds to the number of transitional polygons within each source polygon.

Output product

The output product is delivered in a vector format, presenting number of inhabitants and population

density (inhabitants/sqkm) for each residential spatial unit from the LULC product.




Page 12

Validation

The final product, which is developed by a disaggregated procedure such as the one herein described is

never less accurate than the original source data (JRC Technical Report, 2013). Further detail on the

quality and accuracy of the product is directly linked to the JRC Technical Report (2013), as follows:

‘By disaggregating numerical data from one coarse geometry to a finer geometry, we always gain detail

and approximate ground truth without the risk of deteriorating the source information. The degree to

which the disaggregation approximates reality, however, varies greatly, and it depends chiefly on: 1) the

quality of the ancillary data and 2) the appropriateness of the disaggregation algorithm and its

parameters.’

Reference

Batista e Silva, F., Gallego, J. & Lavalle, C. (2013). A high-resolution grid map for Europe. Journal of

Maps, 9(1), 16-28.

Batista e Silva, F. (DG JRC), Poelman, H. (DG Regional Policy), Martens, V. (DG Regional Policy),

Lavalle, C. (DG JRC). (2013). "Population Estimation for the Urban Atlas Polygons" Joint

Research Center (JRC) Technical Reports.

Kindu, M., Schneider, T., Teketay, D., & Knoke, T. (2015). Drivers of land use/land cover changes in

Munessa-Shashemene landscape of the south-central highlands of Ethiopia. Environmental

monitoring and assessment, 187(7), 452.

Minnesota Population Center. Integrated Public Use Microdata Series, International (IPUMS –

International): Version 6.5. Tanzania – Population and Housing Census 2002. Minneapolis:

University of Minnesota, 2017. http://doi.org/10.18128/D020.V6.5.

United Republic of Tanzania, Ministry of East African Cooperation (MEAC). (2013). 2012

POPULATION AND HOUSING CENSUS TANZANIA. Population Distribution by

Administrative Areas. Downloaded from: http://meac.go.tz/

http://meac.go.tz/




Page 13





Page 14

Annex 2 – Filled Quality Control Sheets




Page 15



Development – Urban Project QA/QC Sheets developed by GAF AG

© 2017 GAF AG All Rights Reserved. Unless otherwise indicated, the templates of these QA/QC pages are copyrighted by GAF AG. No part of these pages, either text or image may be used for any purpose other than use in the EO4SD-Urban Project. Therefore, reproduction, modification, storage in a retrieval system or retransmission, in any form or by any means, electronic, mechanical or otherwise, for reasons other than personal use, is strictly prohibited without prior written permission Page 1

Earth Observation for Sustainable Development - Urban

Quality Assurance and Quality Control Sheets These QA/QC Templates were prepared by GAF AG compliant with ISO 9001:2008 Quality Management System

standards and can only be used by Partners in the current EO4SD-Urban Project.

Project Title: EO4SD-Urban

Project Leader: GAF AG

Service Provider: GAF AG Editor: AB

Client: WB Date: 24.04.2018

Overview of QC-Sheets and Processing Steps Sheet used Sheet filled

in

Requirements

0.1 Requirements Yes Yes

Specifications of Input Data

1.1 List of EO Data Yes Yes

1.2 List of In-situ Data Yes Yes

1.3 List of Ancillary Data Yes Yes

Data Quality Checks

2.1 EO Data Quality Yes Yes

2.2 In-situ Data Quality Yes Yes

2.3 Ancillary Data Quality Yes Yes

Pre-Processing of EO Data

3.1 Geometric Correction Yes Yes

3.1.1 Data Fusion Yes Yes

3.2 Data Processing Yes Yes

Thematic Processing

4.1 Classification Yes Yes

4.2 Intermediate Quality Control of Land Use Data Yes Yes

Accuracy Assessment

5.1 Thematic Accuracy Yes Yes

5.2 Error Matrices Yes Yes

Delivery Checks / Delivery

6.1 Completeness Yes Yes

6.2 Compliancy Yes Yes

Glossary (index numbers in the QA/QC tables refer to the glossary at the end of this document)




Further QC-relevant Documents:

Comments / Characteristics:




0.1 Requirements

Product 1 (28) Urban Land Use/ Land Cover and Change

Abstract

Land Use/Land Cover (LU/LC) information product should contain spatial explicit information on different land use and land cover occurring in both the Core and Peri-Urban areas of the City of Arusha. The Core area has detailed LU/LC nomenclature that is either at Level 3 or 4 whereas the Peri-Urban area LU/LC nomenclature is at an aggregated Level 1 or 2. The input data for the Core area was the Very High Resolution data of WorldView/GeoEye (2015) and Quickbird (2005) and the input data for the Peri-Urban area was Aster (2005) and Sentinel-2 (2015). The LU/LC product is the Baseline Product from which various derived products (such as Green Areas and Informal Settlements) are produced.

Service / Product Specifications

Area Coverage

Country: Tanzania A) Wall-to-wall: yes

City: Arusha Selected Sites: 1

Area km² Core Urban: 211 B) Sampling based: n/a

Peri-Urban: 414

Time Period - Update Frequency

A) Baseline Year(s): B) Update Frequency

2005 and 2015/2016 2 points in time

Comments:

Geographic Reference System

UTM 37S

Mapping Classes and Definitions

Residential (Built-Up Areas)

Built-up areas and their associated land, such as gardens, parks, planted areas and non-surfaced public

areas and the infrastructure, if these areas are not suitable to be mapped separately with regard to the

minimum mapping unit size.

Very High: Average degree of soil sealing: 80 -100% Residential buildings, roads and other artificially

surfaced areas.

High: Average degree of soil sealing: > 50 - 80% Residential buildings, roads and other artificially surfaced

areas.

Medium: Average degree of soil sealing: > 30 - 50% Residential buildings, roads and other artificially

surfaced areas. The vegetated areas are predominant, but the land is not dedicated to forestry or

agriculture.

Low: Average degree of soil sealing: 10 - 30% Residential buildings, roads and other artificially surfaced

areas. The vegetated areas are predominant, but the land is not dedicated to forestry or agriculture.

Super Low: Average degree of soil sealing: <10 % Residential buildings, roads and other artificially

surfaced areas. The vegetated areas are predominant, but the land is not dedicated to forestry or

agriculture. Example: exclusive residential areas with large gardens.

Commercial* warehouses, CBD, shopping malls, markets, other commercial facilities

Industry* Factories, … and associated land

University* University and associated land

Schools* Schools and associated land including sport fields




Hospitals* Hospitals and associated land

Government* Governmental Buildings

Military* Military and associated land

Public Buildings*

“Big” public buildings like churches, bibliotheca, …

Arterial Line Highways, connecting the city with other cities

Collector Line Bigger Connecting roads within the city

Railway Railway facilities including stations, cargo stations and associated land.

Airport Administrative area of airports, mostly fenced. Included are all airport installations: runways, buildings and associated land.

Port Port and associated area

Mining Area and Dump Sites

Open pit extraction sites (sand, quarries) including water surface, if < MinMU, open-cast mines, inland salinas, oil and gas fields;

Construction sites

Spaces under construction or development, soil or bedrock excavations for construction purposes or other earthworks visible in the image.

Land without current use

Areas in the vicinity of artificial surfaces still waiting to be used or re-used. The area is obviously in a transitional position, “waiting to be used”. Waste land, removed former industry areas, (“brown fields”) gaps in between new construction areas or leftover land in the urban context (“green fields”). No actual agricultural or recreational use. No construction is visible, without maintenance, but no undisturbed fully natural or semi-natural vegetation (secondary ruderal vegetation).

Urban Parks

Public green areas for predominantly recreational use such as gardens, zoos, parks, castle parks. Suburban natural areas that have become and are managed as urban parks. Forests or green areas extending from the surroundings into urban areas are mapped as green urban areas when at least two sides are bordered by urban areas and structures, and traces of recreational use are visible.

Recreation facilities

All sports and leisure facilities including associated land, whether public or commercially managed: Golf courses, Sports fields (also outside the settlement area), Camp grounds, Riding grounds, Racecourses, Amusement parks, Swimming resorts etc., Glider or sports airports.

Cemeteries Cemeteries and associated area

Agricultural Area

Cultivated areas non-irrigated or permanently irrigated including rice fields: arable land (annual crops), permanent crops, complex or mixed cultivation, orchards; pasture and meadow under agricultural use, grazed or mechanically harvested.

Forest and Shrub lands

High woody vegetation in natural forests; transitional woodland; low vegetation cover with bushes and shrubs.

Natural Areas Natural area where there is little vegetation and does not serve as construction site.

Bare Soil Natural Area with no vegetation

Wetlands Areas flooded or liable to flooding during a large part of the year by fresh, brackish or standing water with specific vegetation coverage made of low shrub, semi-ligneous or herbaceous species; shallow water areas covered with reed.

Water Visible water areas like lakes, rivers, ponds (natural, artificial).

Comment: * These classes have only to be mapped in 2015. For the mapping of the historic land cover these classes were merged to the class Commercial, Public, Military and Private Units.

Cloud and Cloud Shadow Detection and Removal

Information for the entire AOI is required and therefore clouded areas have to replace by other Earth Observation (EO) data.

Spatial Resolution

n.a. (Product provided as Shapefile)

Minimum Mapping Unit (MMU)

Minimum Mapping Unit is 0.25 ha for the urban area, 0.5 ha for the peri-urban area.

Data Type & Format

Shapefile *.shp and GeoPDF

Bit Depth

n.a.

Class Coding




Class Code Class Name RGB Code

1100 Residential (Built-Up Areas) 255; 0; 0 (main class only)

1211 Commercial 197; 0; 255

1212 Industry 132; 0; 168

1213 University 194; 158; 215

1214 Schools 232; 190; 255

1215 Government 0; 255; 197

1216 Military 115; 115; 0

1217 Hospitals 115; 223; 255

1218 Public Buildings 192; 252; 234

1221 Arterial Line 78; 78; 78

1222 Collector Line 120; 120; 120

1230 Railway 52; 52; 52

1240 Airport 168; 0; 132

1250 Port 0; 168; 132

1310 Mining Area and Dump Sites 115; 76; 0

1320 Construction sites 255; 115; 223

1330 Land without current use 242; 242; 242

1410 Urban Parks 85; 255; 0

1420 Recreation facilities 255; 170; 0

1430 Cemeteries 230; 230; 0

2000 Agricultural Area 255; 235; 175

3100 Forest and Shrub lands 38; 115; 0

3200 Natural Areas 180; 215; 158

3300 Bare Soil 204; 204; 204

4000 Wetlands 76; 0; 115

5000 Water 0; 112; 255

Metadata

Provided as INSPIRE conformant *xml data set, covering at least the mandatory elements.

Service / Product Quality

Thematic Accuracy

Overall Accuracy: >80%

Positional Accuracy

RMSE < 15 m

Delivery Procedure

Service Provision

Online via FTP.

Delivery Date

End of April 2018




1.1 List of EO Data

Sensoren (8) Sentinel-2, Landsat 5, QB-2, PLX

Incoming Date

Acquisition Date

Proc. Level

Path / Row

AOI - City / Region / Country

Spatial Res. No. of Bands

Cloud Cover (Data Provider)

Projection / Spheroid (16)

Data Format (3)

Bit Depth (5)

Header / Metadata (2)

File Name [e.g yymmdd; tbd...]

Sentinel-2

1. S2A_OPER_MSI_L1C_TL_MTI__20160204T095011_A003235_T37MBS

23.03.2017

04.02.2016

1C 37/MBS

Arusha, TZA

10m 4 0.4%

WGS 84 / UTM zone 37S

.tiff 16 Bit

.xml

Aster

2. AST_L1T_00302052006075950_20150513015728_107850

13.15.2015

5.02.2006

L1T 168/ 064

Arusha, TZA

30m 9 0% WGS 84 / UTM zone 37

.hdf 8 bit & 16 bit

.xml

WV-2, GE-1

3.

16SEP17081149-M2AS_R1C1-056358854010_01_P001

31.03.2017

17.09.2016

LV2A R1C1 Arusha, TZA

MUL: 2m; PAN: 0.5m

4 0% UTM 37S .tiff 16 Bit

.xml

16SEP17081149-M2AS_R1C2-056358854010_01_P001

31.03.2017

17.09.2016


MUL: 2m; PAN: 0.5m


.xml

16SEP17081149-M2AS_R1C3-056358854010_01_P001

31.03.2017

17.09.2016


MUL: 2m; PAN: 0.5m


.xml

16SEP17081149-M2AS_R2C1-056358854010_01_P001

31.03.2017

17.09.2016


MUL: 2m; PAN: 0.5m


.xml

16SEP17081149-M2AS_R2C2-056358854010_01_P001

31.03.2017

17.09.2016


MUL: 2m; PAN: 0.5m


.xml

16SEP17081149-M2AS_R2C3-056358854010_01_P001

31.03.2017

17.09.2016


MUL: 2m; PAN: 0.5m


.xml

4. 15DEC30080039-M2AS-056358854010_01_P002

31.03.2017

30.12.2015

LV2A Arusha, TZA

MUL: 2m; PAN: 0.5m


.xml

5. 15JAN04081530-M2AS-056358854010_01_P003

31.03.2017

04.01.2015

LV2A Arusha, TZA

MUL: 2m; PAN: 0.5m

4 4.7e-02%

UTM 37S .tiff 16 Bit

.xml

Quickbird-2

6.

05DEC14082246-M2AS_R1C1-056358854020_01_P001

31.03.2017

05.12.2005


MUL: 2m; PAN: 0.6m

4 0% UTM 37S GeoTIFF

16 Bit

.imd,

.rpb,

.til, xml

05DEC14082246-M2AS_R1C2-056358854020_01_P001

31.03.2017

05.12.2005


MUL: 2m; PAN: 0.6m


16 Bit

.imd,

.rpb,




.til, xml

05DEC14082246-M2AS_R2C1-056358854020_01_P001

31.03.2017

05.12.2005


MUL: 2m; PAN: 0.6m


16 Bit

.imd,

.rpb,

.til, xml

05DEC14082246-M2AS_R2C2-056358854020_01_P001

31.03.2017

05.12.2005


MUL: 2m; PAN: 0.6m


16 Bit

.imd,

.rpb,

.til, xml

7.

05SEP15081836-M2AS_R1C1-056358854020_01_P002

31.03.2017

15.09.2005


MUL: 2m; PAN: 0.6m


16 Bit

.imd,

.rpb,

.til, xml

05SEP15081836-M2AS_R2C1-056358854020_01_P002

31.03.2017

15.09.2005


MUL: 2m; PAN: 0.6m


16 Bit

.imd,

.rpb,

.til, xml




1.2 List of In-situ Data

Dataset 1 (14)

Date of Receipt from Client

Metadata (2)

Reference Mapping Date / Date of Creation

Geographic Area/- City / Region / Country

Area Coverage

Data Type (4)

Data Format (3)


Positional Accuracy (11)

No. of Classes

Thematic Accuracy

Availability of Class Definitions

No In-situ Data Used

Lineage(29):




1.3 List of Ancillary Data

Dataset 1 (14)

Schools Date of Receipt from Client

Metadata (2)



Area Coverage

Data Type (4)

Data Format (3)



No. of Classes

Thematic Accuracy


updated_schools_november_2014.shp

12.04.2016 not

available 2014/n.a. Tanzania (country level) 100% Vector *.shp GCS_WGS_1984 unknown n.a. unknown No

Lineage(29): List of schools in Tanzania.

Source (30):: Received from MaryGrace Weber.

Dataset 2 (14)

Health Facilities


Metadata (2)


Geographic Area / City / Region / Country

Area Coverage

Data Type (4)

Data Format (3)



No. of Classes

Thematic Accuracy

Class Definitions

tanzania_health_facility_registry6322wgs84.shp

12.04.2016 not

available n.a. / n.a. Tanzania (country level) 100% Vector *.shp GCS_WGS_1984 unknown n.a. unknown No

Lineage(29): List of Health Facilities.


Dataset 3 (14)

Landmarks Date of Receipt from Client

Metadata (2)



Area Coverage

Data Type (4)

Data Format (3)



No. of Classes

Thematic Accuracy


ARU_Landmarks.shp 21.12.2016 not

available n.a./n.a.

Arusha (municipal level)

100% Vector *.shp unknown unknown n.a. unknown No

Lineage(29): List of different points of interest in the core urban region, like stadium, bus station, post office, …





Dataset 4 (14)

Roads Date of Receipt from Client

Metadata (2)



Area Coverage

Data Type (4)

Data Format (3)



No. of Classes

Thematic Accuracy


ARU_Roads 21.12.2016 not

available n.a./n.a.

Arusha (municipal level)

100% Vector *.shp unknown unknown n.a. unknown No

Lineage(29): Road network


Dataset 5 (14)

OSM Data Date of Receipt from Client

Metadata (2)


Geographic Area / City / Region / Country

Area Coverage

Data Type (4)

Data Format (3)



No. of Classes

Thematic Accuracy

Class Definitions

OSM_Data_Arusha_Tanz

ania n.a.

not available

n.a. /n.a. Arusha (municipality boundary)

100% Vector *.shp WGS 84 unknown n.a. unknown n.a.

Lineage(29): OSM data were used as basis for the road network as well as for the classification. Railway and Rivers were also extracted and enhanced.

Source (30): https://www.openstreetmap.org/search?query=Arusha#map=11/-3.4363/36.7020




2.1 EO Data Quality

Sensoren(8) Sentinel-2, Landsat 5, QB-2, PLX

Ba

cku

p

Re

ad

ab

ility

(1

)

Ch

eck

He

ad

er

/

Me

tad

ata

(2)

Band Specifications

Pro

ject

ion

/ S

ph

ero

id (1

6)

Sce

ne

Lo

cati

on

Co

mp

lete

ne

ss o

f

Ad

dit

ion

al D

ata

Da

ta F

orm

at

(3)

Bit

De

pth

(5)

Clo

ud

Co

ver

(in

tern

al

che

ck)

Ra

dio

me

try

To

po

gra

ph

y

Dro

pp

ed

Lin

es

/ A

rte

fact

s (7

)

Acc

ep

tan

ce S

tatu

s

Comment

s File Name [e.g yymmdd; tbd...]

No

. o

f B

an

ds

Ba

nd

Re

gis

tra

tio

n

Sp

ect

ral/

Sp

ati

al

Re

solu

tio

n

Sentinel-2


Y - Q Y - V Q Y - V Q Y - V Q Y - V Q Y - Q Y - V Y - V Y - V Q Y - Q Y - Q

Y - V Q

Y - V Q

Y - V

N - V Q

Yes none

Aster

2. AST_L1T_00302052006075950_20150513015728_107850


Y - V Q

Y - V Q

Y - V

N - V Q

Yes none

GeoEye-1/Worldview-2

3. 16SEP17081149-M2AS-056358854010_01_P001


Y - V Q

Y - V Q

Y - V

N - V Q

Yes none

4. 15DEC30080039-M2AS-056358854010_01_P002


Y - V Q

Y - V Q

Y - V

N - V Q

Yes none

5. 15JAN04081530-M2AS-056358854010_01_P003


Y - V Q

Y - V Q

Y - V

N - V Q

Yes none

Quickbird-2

6. 05DEC14082246-M2AS-056358854020_01_P001


Y - V Q

Y - V Q

Y - V

N - V Q

Yes none

7. 05SEP15081836-M2AS-056358854020_01_P002


Y - V Q

Y - V Q

Y - V

N - V Q

Yes none




2.2 In-situ Data Quality

Dataset 1 (14)

Ba

cku

p

Re

ad

ab

ility

(1)

Ch

eck

He

ad

er

/

Me

tad

ata

(2)

Ext

en

t

Pro

ject

ion

/ S

ph

ero

id

(16

)

Sp

ati

al R

eso

luti

on

Da

ta F

orm

at

(3)

Bit

De

pth

(5)

Lo

cati

on

Co

mp

lete

ne

ss

Ge

om

. M

isa

lign

me

nt

Pla

usi

bili

ty

Dro

pp

ed

Lin

es

/ A

rte

fact

s (7

)

Acc

ep

tan

ce S

tatu

s

Comments





2.3 Ancillary Data Quality

Dataset 1 (14) LU Map

Re

ad

ab

ility

(1)

Ch

eck

He

ad

er

/Me

tad

ata

(2)

Da

ta C

ove

rage

M

atc

he

s S

erv

ice

A

rea

(%

)

Pro

ject

ion

/ S

ph

ero

id

, E

PS

G(1

6)

Sp

ati

al R

eso

luti

on

a

nd

un

it (

e.g

. m

, k

m)

Da

ta F

orm

at

(3)

Bit

De

pth

(5)

Co

mp

lete

ne

ss

(Ve

cto

r: A

ttri

bu

te

Ta

ble

; R

ast

er:

Th

em

ati

c V

alu

es)

Ge

om

. M

isa

lign

me

nt

Dro

pp

ed

Lin

es

/

Art

efa

cts

(E

O d

ata

o

nly

)(7

)

Uti

lity

for

Cu

rre

nt

Pro

ject


updated_schools_november_2014.shp

☒ Yes

☐ No

☐ Complete

(INSPIRE/ISO19119)

☐ Incomplete

☒ not available

☒ Yes

☐ No

☐ Unknown

If no: xxx%

☒ Correct

☐ Incorrect

☐ Unknown

EPSG: 32749

n.a. *.shp n.a. ☒ Complete

☐Incomplete

☐ No

☐ Yes

☒ Unknown

If yes: XX m

☐ No

☐ Yes

☒ n.a.

If yes: xxx %

☐ None

☒ Partial

☐ Full

Comments: Accuracy of the dataset is unknown. Some points show schools, others not.

Dataset 2,3,4 (14)

Re

ad

ab

ility

(1)

Ch

eck

He

ad

er

/Me

tad

ata

(2)

Da

ta C

ove

rage

M

atc

he

s S

erv

ice

Are

a

(%)

Pro

ject

ion

/ S

ph

ero

id

EP

SG

(16

)

Sp

ati

al R

eso

luti

on

a

nd

un

it (

e.g

. m

, k

m)

Da

ta F

orm

at

(3)

Bit

De

pth

(5)

Co

mp

lete

ne

ss

(Ve

cto

r: A

ttri

bu

te

Ta

ble

; R

ast

er:

Th

em

ati

c V

alu

es)

Ge

om

. M

isa

lign

me

nt

Dro

pp

ed

Lin

es

/

Art

efa

cts

(E

O d

ata

o

nly

)(7

)

Uti

lity

for

Cu

rre

nt

Pro

ject


tanzania_health_facility_registry6322wgs84.shp

☒ Yes

☐ No

☐ Complete

(INSPIRE/ISO19119)

☐ Incomplete

☒ not available

☐ Yes

☐ No

☒ Unknown

If no: %

☒ Correct

☐ Incorrect

☐ Unknown

EPSG: 32749


☐Incomplete

☐ No

☐ Yes

☒ Unknown

If yes: XX m

☐ No

☐ Yes

☒ n.a.

If yes: xxx %

☐ None

☒ Partial

☐ Full

ARU_Landmarks.shp ☒ Yes

☐ No

☐ Complete

(INSPIRE/ISO19119)

☐ Incomplete

☒ not available

☐ Yes

☐ No

☒ Unknown

If no: %

☒ Correct

☐ Incorrect

☐ Unknown

EPSG: 32749

n.a. *.shp n.a. ☐ Complete

☒Incomplete

☐ No

☐ Yes

☒ Unknown

If yes: XX m

☐ No

☐ Yes

☒ n.a.

If yes: xxx %

☐ None

☒ Partial

☐ Full

ARU_Roads ☒ Yes

☐ No

☐ Complete

(INSPIRE/ISO19119)

☐ Incomplete

☒ not available

☐ Yes

☐ No

☒ Unknown

If no: %

☒ Correct

☐ Incorrect

☐ Unknown

EPSG: 32749


☒Incomplete

☐ No

☐ Yes

☒ Unknown

If yes: XX m

☐ No

☐ Yes

☒ n.a.

If yes: xxx %

☐ None

☒ Partial

☐ Full

OSM_Data_Arusha_Tanzania ☒ Yes

☐ No

☐ Complete

(INSPIRE/ISO19119)

☐ Incomplete

☒ not available

☐ Yes

☐ No

☒ Unknown

If no: %

☒ Correct

☐ Incorrect

☐ Unknown

EPSG: 32749


☒Incomplete

☐ No

☐ Yes

☒ Unknown

If yes: XX m

☐ No

☐ Yes

☒ n.a.

If yes: xxx %

☐ None

☒ Partial

☐ Full




Comments:

tanzania_health_facility_registry6322wgs84.shp:

Accuracy of the data is unknown. Some points show schools, others not. ARU_Landmarks.shp:

Accuracy of the data is unknown.

ARU_Roads.shp:

Accuracy of the data is unknown.

OSM_Data_Arusha_Tanzania:

Were used in addition to the Road layers received from the city of Semarang. They were visually checked and corrected




3.1 Geometric Correction

Sensoren (8) Sentinel-2, Landsat 5, QB-2, PLX

Pro

cess

ing D

ate

AO

I

Cit

y /

Re

gio

n /

Co

un

try

Pro

ject

ion

/ S

ph

ero

id (1

6)

No

. &

RM

S (

m)

of

GC

Ps

(17

)

No

. &

RM

S (

m)

of

TP

s (1

8)

No

. &

RM

S (

m)

of

CP

s (1

9)

Dig

ita

l E

leva

tio

n M

od

el

(DE

M)

Mo

de

l /

Alg

ori

thm

(13

)

Re

sam

plin

g

Me

tho

d (2

0)

Va

lida

tio

n R

ep

ort

s

Acc

ep

tan

ce S

tatu

s

Output File Name [e.g yymmdd; tbd...] N

o. File Name [e.g yymmdd; tbd...]

Sentinel-2


12.04.2017

Arusha TZA

UTM37S/ WGS84

N/A N/A N/A N/A N/A N/A N/A Yes

S2A_OPER_MSI_L1C_TL_MTI__20160204T095011_A003235_T37MB

S

Aster

2.

AST_L1T_00302052006075950_20150513015728_107850_SWIR_456789_utm37_toa

12.04.2017

Arusha TZA

UTM37S/ WGS84


AST_L1T_00302052006075950_20150513015728_107850_SWIR_456789_utm37_toa

AST_L1T_00302052006075950_20150513015728_107850_VNIR123_utm37_toa

12.04.2017

Arusha TZA

UTM37S/ WGS84


AST_L1T_00302052006075950_20150513015728_107850_VNIR123_utm37_toa


3. 15DEC30080039-M2AS-056358854010_01_P002_toa_PSH

12.04.2017

Arusha TZA

UTM37S / WGS84

MGCPs 11;

RMSE x1,17; y

1,07

4; RMSE 0,61; y 0,67

N/A SRTM

30 0 order polynom

CC Yes Yes

o15DEC30080039-M2AS-056358854010_01_P002_toa_PSH

4. 16SEP17081149-M2AS-056358854010_01_P001_toa_PSH

12.04.2017

Arusha TZA

UTM37S / WGS84

MGCPs 11;

RMSE x1,17; y

1,07

4; RMSE 0,61; y 0,67

N/A SRTM 30

0 order polynom

CC Yes Yes

o16SEP17081149-M2AS-056358854010_01_P001_toa_PSH

5. 15JAN04081530-M2AS-056358854010_01_P003_toa_PSH

12.04.2017

Arusha TZA

UTM37S / WGS84

MGCPs 11;

RMSE

4; RMSE 0,61; y 0,67

N/A SRTM 30

0 order polynom

CC Yes Yes

o15JAN04081530-M2AS-056358854010_01_P003_toa_PSH




x1,17; y 1,07

Quickbird-2

6. 05DEC14082246-M2AS-056358854020_01_P001_toa_PSH

12.04.2017

Arusha TZA

UTM37S / WGS84

MGCPs 15;

RMSE x1,01; y0,79

N/A N/A SRTM 30

Affin, 1st order poynom

CC Yes Yes

o05DEC14082246-M2AS-056358854020_01_P001_toa_PSH

7. 05SEP15081836-M2AS-056358854020_01_P002_toa_PSH

12.04.2017

Arusha TZA

UTM37S / WGS84

MGCPs 15;

RMSE x1,01; y0,79

N/A N/A SRTM 30


CC Yes Yes

o05SEP15081836-M2AS-056358854020_01_P002_toa_PSH




3.1.1 Data Fusion

Dataset 4 (14) Input A Input B Output

Acc

ep

tan

ce S

tatu

s

Comments

Filename 16SEP17081149-M2AS-

056358854010_01_P001 16SEP17081149-P2AS-

056358854010_01_P001 16SEP17081149-M2AS-

056358854010_01_P001_toa_PSH

none

Sensor GeoEye-1 GeoEye-1 GeoEye-1

Method PCI / Pansharp2 PCI / Pansharp2 PCI / Pansharp2

Spatial Resolution / (MS/Pan)

2m 0,5m 0,5m 0,5m 0,5m MS

Band Combination BGR NIR PAN BGR NIR

Data Format (3) & Bit Depth (5)

*.tif 16 Bit u *.tif 16 Bit u *.pix 16 Bit u Yes


Acc

ep

tan

ce S

tatu

s

Comments

Filename 15DEC30080039-M2AS-

056358854010_01_P002 15DEC30080039-P2AS-

056358854010_01_P002 15DEC30080039-M2AS-

056358854010_01_P002_toa_PSH

none




2m 2m 0,5m Pan 0,5m MS





Acc

ep

tan

ce S

tatu

s

Comments

Filename 15JAN04081530-M2AS-

056358854010_01_P003 15JAN04081529-P2AS-

056358854010_01_P003 15JAN04081530-M2AS-

056358854010_01_P003_toa_PSH

none

Sensor Worldview-2 Worldview-2 Worldview-2



2m MS 0,5m Pan 0,5m MS








Acc

ep

tan

ce S

tatu

s

Comments


056358854020_01_P001 05DEC14082245-P2AS-

056358854020_01_P001 05DEC14082246-M2AS-

056358854020_01_P001_toa_PSH

none

Sensor Quickbird-2 Quickbird-2 Quickbird-2








Acc

ep

tan

ce S

tatu

s

Comments


056358854020_01_P002 05SEP15081836-P2AS-

056358854020_01_P002 05SEP15081836-M2AS-

056358854020_01_P002_toa_PSH

none











3.2 Data Processing

Sensoren (8)

Landsat 5, QB-2, PLX Atmospheric Correction

Radiometric Processing (15)

Topographic Normalisation

OTHER Calculation:

Acc

ep

tan

ce S

tatu

s

Ba

cku

p

Comment

No.

File Name [e.g yymmdd; tbd...] Processing

Date

pro

cess

ed

Software / Method

pro

cess

ed

Software / Method

pro

cess

ed

Software / Method

pro

cess

ed

Software / Method

Sentinel-2

S2A_OPER_MSI_L1C_TL_MTI__20160204T095011_A003235_T37MBS

10.04.2017 No N/A No N/A No N/A No N/A Yes Yes none

Aster

2. AST_L1T_00302052006075950_20150513015728_107850_VNIR123_utm37

12.04.2017 No N/A No N/A No N/A Yes GAFmap /

TOA Yes No none

GeoEye1/Worldview-2

3. 16SEP17081149-M2AS-056358854010_01_P001

12.04.2017 No N/A No N/A No N/A Yes GAFmap / TOA

Yes No none

4. 15DEC30080039-M2AS-056358854010_01_P002


Yes No none

5. 15JAN04081530-M2AS-056358854010_01_P003


Yes No none

Quickbird-2

6. 05DEC14082246-M2AS-056358854020_01_P001


Yes No none

7. 05SEP15081836-M2AS-056358854020_01_P002


Yes No none




4.1 Classification

Sensors (8) Sentinel-2, Aster, WorldView, GeoEye, QB

Processing Date

Cloud Masking Thematic

Classification Manual

Enhancement Mosaicking

Acc

ep

tan

ce S

tatu

s

Ba

cku

p

Comment No.


pro

cess

ed

Software / Method

pro

cess

ed

Software / Method

pro

cess

ed

Software /

Method

pro

cess

ed

Software / Method


09.06.2017 no No clouds Yes

eCognition/semi-automatic approach

Yes

eCognition and ArcGIS 10 / Visual Interpretation

Yes

ArcGIS 10 / Mosaic Dataset with manual neatline selection

Yes Yes

2. AST_L1T_00302052006075950_20150513015728_107850_VNIR123_utm37

07.08.2017 no View clouds No NA No NA No

Only used to assist visual interpretatio

n

3.

16SEP17081149-M2AS-056358854010_01_P001 15DEC30080039-M2AS-056358854010_01_P002 15JAN04081530-M2AS-056358854010_01_P003


GAFMap and ArcMap/visual mapping

Yes

GAFMap and ArcMap/Visual interpretation

Yes


Yes Yes

4.

05DEC14082246-M2AS-056358854020_01_P001 05SEP15081836-M2AS-056358854020_01_P002



Yes


Yes


Yes Yes


Development – Urban

Project QA/QC Sheets developed by GAF AG


4.2 Intermediate Quality Control of LCLU Data

No.

x = ok (checked/performed) o = not performed

n/a = not applicable

Co

mp

lian

t

Each item was checked during production and if necessary the source of error was identified and

corrected.

Description of QC item Comment

1 Data in predefined coordinate system?

(e.g. UTM WGS84 system) x

2 Do all features have attributes? x

3

All feature attributes are valid (attribute

value range, flags, correct comments)?

1) Do all features have a valid class

name?

x

4 All feature attributes present and filled

in? x

5 Data coherence at the border of the

urban area checked? x

6 Positional Accuracy of the features

according to User Specifications? x

7 All feature geometries valid? x

8

Vertices of feature geometries adjusted?

(e.g. missing, duplicated/ very close

vertices)

x

9 Multipart features resolved (Ring loops)? x

10 No data gaps? (Clean Gaps only if there

are over ~200 gaps)

11 No data overlaps? (Clean Overlaps only if

there are over ~200 gaps) x

12

Feature geometries optimized (e.g.

spikes, cutbacks minimized) (Find acute

angles, remove angles < 15°)

x

13 Minimum Mapping Area checked? (Urban

0.25ha, peri-urban 0.5ha) x

14 Minimum Mapping Width checked? x

15 Data checked for unnecessary polygon

boundaries? (after final dissolve) x

16 Service Area completely covered by

requested data (no data gaps)? x

17 No data outside delivery AoI? x

18

Add two new columns with the code and

the name of the class.(Name them:

code_year and name_year of the actual

classification”

19

Feature attribution and feature relations

plausible? (Plausibility Check ) --> Change

product

x

20 Calculation of area [AREA_km2] adjusted

to final data? x

21 Re-calculation of [ID] (FID + 1)? x

22

LCLU product file naming and versioning

according to specifications?

(City_program_product_referenceDate_C

oordinateSystem)

x




5.1 Thematic Accuracy

No. Product Name

Processing Date

Area Coverage

No

. o

f C

lass

es

Re

fere

nce

Da

ta

Sa

mp

le S

ize

Sa

mp

ling U

nit

(21

)

Sa

mp

ling

De

sign

(22

)

Sa

mp

le E

xclu

sio

n

Cri

teri

a (2

3)

Th

em

ati

c A

ccu

racy

(2

4)

Acc

ep

tan

ce S

tatu

s

Comment

1.

Land Cover Map 2015

14.08.2017 Full AOI (urban and peri-urban)

Mapped. 22

classes

VHR data sources for urban area. Bing Maps by Microsoft and Google Imagery for Peri-urban area.

215 per strata (Level 1 classes)

Point Stratified random sampling None, all used 88.8% Y

Four classes did not occur: ‘1216 Military’; ‘1230 Railway’; ‘1250 Port’ and ‘3300 Bare Soil’




5.2 Error Matrices

Class Name (columns = Ground Truth; Rows = Mapped Class)

Re

sid

en

tia

l

Co

mm

erc

ial

Ro

ad

s

Air

po

rt

Min

ing

Co

nst

ruct

ion

La

nd

wit

ho

ut

curr

en

t u

se

Urb

an

Pa

rks

Re

cre

ati

on

al

Fa

cilit

ies

Ce

me

tery

Agri

cult

ure

Fo

rest

Na

tura

l A

rea

s

Wa

ter

User Accuracy and Confidence Interval at 95% Confidence

Level

Class ID 11 121 122 124 131 132 133 141 142 143 2 31 32 5

Totals

Residential 11 139 9 0 0 0 0 0 0 0 0 22 6 2 0 178 78.1 % ± 6.4 Commercial 121 1 13 0 0 2 0 0 0 16 81.3 % ± 22.3

Roads 122 0 0 147 0 0 0 0 0 0 0 0 3 0 0 150 98.0 % ± 2.6 Airport 124 0 0 0 1 0 0 0 0 0 0 0 0 0 0 1 100.0 % ± n.a. Mining 131 0 0 0 0 66 0 0 0 0 0 0 0 0 0 66 100.0 % ± 0.7

Construction 132 0 0 0 0 0 45 0 0 0 0 0 0 3 0 48 93.8 % ± 7.9 Land without current use 133 0 0 0 0 0 0 6 0 0 0 1 0 0 0 7 85.7 % ± 33.1

Urban Parks 141 0 1 0 0 0 0 0 11 0 0 0 0 0 0 12 91.7 % ± 19.8 Recreational Facilities 142 0 0 0 0 0 0 0 0 1 0 0 0 0 0 1 100.0 % ± n.a

Cemetery 143 0 0 0 0 0 0 0 0 0 3 0 0 0 0 3 100.0 % ± 16.7 Agriculture 2 1 0 0 0 0 0 0 0 0 0 192 18 4 0 215 89.3 % ± 4.4

Forest 31 7 0 0 0 0 0 0 2 1 0 11 154 7 0 182 84.6 % ± 5.5 Natural Areas 32 1 0 0 0 0 0 0 0 0 0 3 0 29 0 33 87.9 % ± 12.7

Water 5 0 0 0 0 0 0 0 0 0 0 0 0 0 29 29 100.0 % ± 1.7

Totals 149 23 147 1 66 45 6 13 2 3 231 181 45 29 941

Producer Accuracy and Confidence Interval at 95% Confidence Level

93

.3%

±

4.4

56

.5 %

±

22

.4

10

0.0

%

±0

.3

10

0.0

%

±n

.a.

10

0.0

%

± 0

.8

10

0,0

0%

±1

.1

10

0.0

% ±

8

.3

84

,62

%

± 2

3.5

50

,00

%

±n

.a

10

0,0

0%

±

n.a

83

,12

%

±5

.1

85

,08

%

±5

.5

64

,44

%

±1

5.1

10

0.0

%

±1

.7 Overall Accuracy: 88.8%

Confidence Interval 86.8% - 90.9%




6.1 Completeness

INPUT DATA

No.

Item AOI Coverage [km²]

Area Coverage [km²]

Completeness of Coverage (25)

No. of Scenes Scenes used in Production

Completeness of Verification (27)

Metadata Comments

1. HR EO Data

Arusha

414 km² 100% 2 2 100% Yes none

2. VHR EO Data 211 km² 100% 5 5 100% Yes none

3. In-situ Data N/A N/A N/A N/A N/A N/A none

4. Ancillary Data N/A N/A N/A N/A N/A N/A none

PRODUCTS

No.

Product (28) AOI Coverage [km²]

Product Coverage [km²]


Completeness of Classification (26)

Unclassifiable Area [%]


Metadata Comments

1. Urban and Peri Urban Areas 2005 and 2015

625 km² 625 km² 100% 100% 0% 100% Yes none




6.2 Compliancy

Product 1 (28) EO4SD_Arusha_WB_LULC_2005_2015

Abstract

Land Use/Land Cover (LU/LC) information product contains spatial explicit information on different land use and land cover occurring in both the Core and Peri-Urban areas of the City of Arusha. The Core area has detailed LU/LC nomenclature that is either at Level 3 or 4 whereas the Peri-Urban area LU/LC nomenclature is at an aggregated Level 1 or 2. The input data for the Core area was the Very High Resolution data of WorldView/GeoEye (2015) and Quickbird (2005) and the input data for the Peri-Urban area was Aster (2005) and Sentinel-2 (2015). The LU/LC product is the Baseline Product from which various derived products (such as Green Areas and Informal Settlements) are produced.


Area Coverage

Requirements Achieved Specifications Compliancy Comments

Country 1: Tanzania Country 1: Tanzania Yes ---

Country 2: N/A Country 2: N/A --- ---



A) Wall-to-wall: Arusha city A) Wall-to-wall: Arusha city Yes --- Core (211 km²) and Peri-Urban

area (414 km²) Full coverage of AOIs, Core and Peri-Urban area

Area of Interest. 625 km², defined by the national user.

B) Sampling based: B) Sampling based: --- ---

N/A N/A



Baseline Year(s): Baseline Year(s): Yes

See EO data acquisition dates. 2005 and 2015 (+/- 1 years) 2005 and 2015

B) Update Frequency B) Update Frequency Yes --

No update No update



EPSG:32737; WGS84 / UTM 37S

EPSG:32737; WGS84 / UTM 37S Yes --


Requirements Achieved Specifications Compliancy

26 Main classes (see 0.1 Requirements sheet) All classes were mapped according to their definition.

Yes

Cloud and Shadow Detection and Removal


Cloud and shadow covered areas shall be removed.

Full AoI coverage with spatial explicit LULC cover information

Yes ---




Spatial Resolution


Na Na Na Only applicable for raster

data



0.25 ha and 0.5 ha MMU of 0.25 ha and 0.5 ha Yes ---

Data Type


Vector data yes Yes ---

GeoPDF None were set Na ---

Bit Depth


None Na Na ---

Data Format


*.shp Shapefile Shapefile Yes ---

(open cross-platform format)

Class Coding (Raster Data Only)


Na Na ---

Metadata


Metadata ISO compliant INSPIRE compliant and attached to

product Yes ---


Thematic Accuracy


Overall Accuracy: >80% 88% Yes ---

Positional Accuracy


RMSE < 15 m 15m Y ---

Delivery Procedure

Service Provision





online via FTP Uploaded to FTP Yes ---

Delivery Date


End of August 2017 Third Delivery April 2018 Yes ---




Glossary

Quality Checks

Prefix Suffix Explanation

Y (Yes) - V Visually checked

N (No) - Q Quantitatively/qualitatively checked

N/A - V Q Visually and quantitatively/qualitatively checked

- N/A Not applicable

Readability (1) Check readability of all required input data. Can the data be stored again?

Header / Metadata (2) Check Image Header Information and/or Metadata for completeness / distinctive features.

Data Format (3) For digital data, please give file format (e.g. *.tiff, *.shp).

Data Type (4) Please specify the type of the data (e.g. raster, vector or analogue).

Bit Depth (5) Please give pixel depth and sign of raster data (e.g. 8 bit unsigned integer).

Dynamic Range (6) Check dynamic range of all image bands. Visual check of dynamic range should be accompanied by histograms and statistics.

Dropped Lines & Artefacts (7) Check Image for dropped lines and other artefacts. If such occurs, please give description of extent and influence in the Comments section.

Sensor 1, 2, … (8) Please delete / add additional sensor sections as necessary.

Purpose (9)

The purpose and use of the given auxiliary or reference data should be stated:

ORHTOrectification, GEOmetric correction, REFerence, VERification source, POSitional ACCuracy assessment, THEmatic ACCuracy assessment.

Sampling Methodology (10)

Methodology of in situ or reference/auxiliary data sampling scheme should be outlined. In case of sample plots, also state how the plot positions have been determined (e.g. from GPS measurements, topographic maps, terrestrial triangulation, EO data, etc.) and how the sampling grid was established.


Positional accuracy of collected in situ data should be given. For reference/auxiliary data it MUST be given. If unknown, the data’s use must be explained. For DEM or other data with 3D information please specify both vertical and horizontal Positional Accuracy. For analogue data (e.g. maps) try to give approximate accuracy related to mapping scale.

Completeness (12) Data should be checked for spatial/temporal/content gaps.

Model / Algorithm (13)

Give the name of the software and its version. Specify the software module/algorithm used for: a) geometric correction, e.g., Polynomial and its degree, Rational Functions, Thin Plate Spline, etc. b) classification, e.g., ISODATA, Maximum Likelihood, Neural Networks, etc.

Dataset 1, 2, … (14) Please delete / add additional dataset sections as necessary.

Radiometric Processing (15) State whether (and which) radiometric processing was applied (e.g., contrast enhancement, histogram matching, bundle block adjustment, radiometric normalisation, filtering, etc.).

Projection; Spheroid / Ellipsoid (16)

Always specify completely, i.e. at minimum the Projection (+Zone, if applicable), Spheroid / Ellipsoid, Map Datum. Give additional information if necessary to unambiguously define the reference system.




Ground Control Points (GCPs) (17)

Give the number, distribution and RMS of used Ground Control Points (as average per scene) as obtained from the Geometric Correction, in meters [m]. Optionally, also RMSx and RMSy may be given. Distribution of GCPs should be attached as a snapshot, or described in the Comments section.

Tie Points (TPs) (18)

In case of mosaicking, give the number of used Tie Points and their total RMS (as average per scene) as obtained from the Geometric Correction, in meters [m]. Optionally, also RMSx and RMSy may be given. Distribution of TPs should be attached as a snapshot, or described in the Comments section.

Check Points (CPs) (19)

The real measure of positional accuracy and the only measure which should be examined as to its Acceptable Range. Give the independent Check Points’ total RMS (as average per scene) as obtained from the Geometric Correction, in meters [m]. Optionally, also RMSx and RMSy may be given. Distribution of CP's should be attached as a snapshot, or described in the Comments section.

Resampling Method (20) Specify, if / which resampling algorithm has been used (e.g. NN=Nearest Neighbour, BIL=Bilinear Interpolation, CC=Cubic Convolution).

Sampling unit (21) Specify sampling unit of Accuracy Assessment as POINT, FRAME, POLYGON and how it is treated (e.g. pixel center, polygon centre, etc.).

Sampling Design (22) SYST=Systematic, RAND=Random, STRAT=Stratified, SBCLASS=Stratified by class, SBAREA=Stratified by area

Sample exclusion criteria (23)

Describe which criteria you apply for sampling point selection resp. exclusion of certain points. For example, if the point is too close to a class boundary (less than 1 pixel), it is excluded. If the selected sample point is not representative of the class , it is excluded. For these reasons, it is recommended to oversample by 10% to compensate for sample point exclusion.

Thematic Accuracy (24)

Provide a detailed description of the Accuracy Assessment results in the form of error matrices showing commission and omission errors, user’s , producer’s and overall accuracies and other measures of Thematic Accuracy, as the confidence level (usually fixed at 95%) and the respective confidence interval, at least for the overall accuracy. If classification is done in a phased approach, e.g. if Forest Area and subsequently Forest Type are mapped, independent reports have to be produced.

Completeness of Coverage (25) State whether the coverage is limited to a subset, or portion of the final product.


State whether classification was constrained to a subset, or portion of the final product.


State whether verification applied to lineage, positional, or thematic accuracy is constrained to a subset, or portion of the final product.

Product (28) Please delete / add additional product sections as necessary.

Lineage (29) Definition (INSPIRE 2015): Lineage is “a statement on process history and/or overall quality of the spatial data set. Where appropriate it may include a statement whether the data set has been validated or quality assured, whether it is the official




version (if multiple versions exist), and whether it has legal validity. The value domain of this element is free text.”

Source(30) Definition (INSPIRE, 2015): “This is the description of the organisation responsible for the establishment, management, maintenance or distribution of the resource. This description shall include: name of the organisation and contact email address.”









Service Provider: GAF Editor: A. Broszeit


Product: Urban Green Areas

Overview of QC-Sheets and Processing Steps Sheet used Sheet filled

in

Requirements






Data Quality Checks








Thematic Processing


Accuracy Assessment






Glossary (index numbers in the QA/QC tables refer to the glossary at the end of this document) Further QC-relevant Documents:








0.1 Requirements

Product 1 (28) Urban Green Areas 2005 and 2015

Abstract

Urban Green Areas information product contains spatial explicit information on green areas within the urban extent. Green areas include linear green features such as river alignments, hedges and trees, as well as public parks, private gardens, forested areas, etc. Urban Green Areas & Change dataset is based on Very High Resolution (VHR) satellite imagery by means of automated classification processing techniques.


Area Coverage

Country: Tanzania A) Wall-to-wall: n/a

City: Arusha Selected Sites: Area of Interest defined with end-users





Comments: None


WGS84 UTM Zone 37S


Single date

0 Non-urban green area

1 Urban green area

Change product

0 Non-urban green area

1 Permanent urban green area

2 Loss of urban green area

3 New urban green area


n/a

Spatial Resolution

1 m


1 m²

Data Type & Format

Raster data in GEOTIFF




Bit Depth

8 bit for *.tif raster files

Class Coding


Single date

0 Non-urban green area 240/240/240

1 Urban green area 56/168/0

Change product

0 Non-urban green area 255/245/235

1 Permanent urban green area 56/168/0

2 Loss of urban green area 255/100/100

3 New urban green area 152/230/0

Metadata

ISO standard 19115 "Metadata" and ISO 19139 "XML Scheme Implementation"


Thematic Accuracy

> 80%

Positional Accuracy

3 m (CE90)

Delivery Procedure

Service Provision

Online via FTP

Delivery Date

End of April 2018




1.1 List of EO Data

Sensoren (8) GE-1, WV-2, QB-2

Incoming Date

Acquisition Date

Proc. Level

Path / Row





Data Format (3)

Bit Depth (5)



WV-2, GE-1

1.

16SEP17081149-M2AS_R1C1-056358854010_01_P001

31.03.2017

17.09.2016


MUL: 2m; PAN: 0.5m


.xml

16SEP17081149-M2AS_R1C2-056358854010_01_P001

31.03.2017

17.09.2016


MUL: 2m; PAN: 0.5m


.xml

16SEP17081149-M2AS_R1C3-056358854010_01_P001

31.03.2017

17.09.2016


MUL: 2m; PAN: 0.5m


.xml

16SEP17081149-M2AS_R2C1-056358854010_01_P001

31.03.2017

17.09.2016


MUL: 2m; PAN: 0.5m


.xml

16SEP17081149-M2AS_R2C2-056358854010_01_P001

31.03.2017

17.09.2016


MUL: 2m; PAN: 0.5m


.xml

16SEP17081149-M2AS_R2C3-056358854010_01_P001

31.03.2017

17.09.2016


MUL: 2m; PAN: 0.5m


.xml

2. 15DEC30080039-M2AS-056358854010_01_P002

31.03.2017

30.12.2015

LV2A Arusha, TZA

MUL: 2m; PAN: 0.5m


.xml

3. 15JAN04081530-M2AS-056358854010_01_P003

31.03.2017

04.01.2015

LV2A Arusha, TZA

MUL: 2m; PAN: 0.5m

4 4.7e-02%


.xml

Quickbird-2

4.

05DEC14082246-M2AS_R1C1-056358854020_01_P001

31.03.2017

05.12.2005


MUL: 2m; PAN: 0.6m


16 Bit

.imd,

.rpb,

.til, xml

05DEC14082246-M2AS_R1C2-056358854020_01_P001

31.03.2017

05.12.2005


MUL: 2m; PAN: 0.6m


16 Bit

.imd,

.rpb,

.til, xml

05DEC14082246-M2AS_R2C1-056358854020_01_P001

31.03.2017

05.12.2005


MUL: 2m; PAN: 0.6m


16 Bit

.imd,

.rpb,

.til, xml

05DEC14082246-M2AS_R2C2-056358854020_01_P001

31.03.2017

05.12.2005


MUL: 2m; PAN: 0.6m


16 Bit

.imd,

.rpb,

.til, xml




5.

05SEP15081836-M2AS_R1C1-056358854020_01_P002

31.03.2017

15.09.2005


MUL: 2m; PAN: 0.6m


16 Bit

.imd,

.rpb,

.til, xml

05SEP15081836-M2AS_R2C1-056358854020_01_P002

31.03.2017

15.09.2005


MUL: 2m; PAN: 0.6m


16 Bit

.imd,

.rpb,

.til, xml





Dataset 1 (14) Incoming Date Acquisition Date AOI - City / Region / Country

Data Type (4)


No. of Sample Plots

Sampling Design (22)


Purpose (9)


Lineage:





Dataset 1 (14)

LULC Date of Receipt from Client

Metadata (2)



Area Coverag

e

Data Type (4)

Data Format (3)



No. of Classes

Thematic

Accuracy


EO4SD-Urban Land Use/Land Cover Baseline Product

2017-07-30 Yes

(INSPIRE) 2006 and 2016

Arsuha (urban area) 100% Vector *.shp EPSG:32736, WGS 84 / UTM zone 36S

< 3 m 24 88% Yes

Lineage(29): /

Source (30):: GAF AG




2.1 EO Data Quality

Sensoren(8)

WV, GE GE-1, WV-2, QB-2

Ba

cku

p

Re

ad

ab

ility

(1

)

Ch

eck

He

ad

er

/ M

eta

da

ta

(2)

Band Specifications

Pro

ject

ion

/ S

ph

ero

id (1

6)

Sce

ne

Lo

cati

on

Co

mp

lete

ne

ss o

f

Ad

dit

ion

al D

ata

Da

ta F

orm

at

(3)

Bit

De

pth

(5)

Clo

ud

Co

ver

(in

tern

al

che

ck)

Ra

dio

me

try

To

po

gra

ph

y

Dro

pp

ed

Lin

es

/ A

rte

fact

s (7

)

Acc

ep

tan

ce S

tatu

s

Comment


No

. o

f B

an

ds

Ba

nd

Re

gis

tra

tio

n

Sp

ect

ral/

Sp

ati

al

Re

solu

tio

n


1. 16SEP17081149-M2AS-056358854010_01_P001


Y - V Q

Y - V Q

Y - V

N - V Q

Yes none

2. 15DEC30080039-M2AS-056358854010_01_P002


Y - V Q

Y - V Q

Y - V

N - V Q

Yes none

3. 15JAN04081530-M2AS-056358854010_01_P003


Y - V Q

Y - V Q

Y - V

N - V Q

Yes none

Quickbird-2

4. 05DEC14082246-M2AS-056358854020_01_P001


Y - V Q

Y - V Q

Y - V

N - V Q

Yes none

5. 05SEP15081836-M2AS-056358854020_01_P002


Y - V Q

Y - V Q

Y - V

N - V Q

Yes none





Dataset 1 (14) No Data used

Ba

cku

p

Re

ad

ab

ility

(1)

Ch

eck

He

ad

er

/ M

eta

da

ta (2

)

Ext

en

t

Pro

ject

ion

/ S

ph

ero

id

(16

)

Sp

ati

al R

eso

luti

on

Da

ta F

orm

at

(3)

Bit

De

pth

(5)

Lo

cati

on

Co

mp

lete

ne

ss

Ge

om

. M

isa

lign

me

nt

Pla

usi

bili

ty

Dro

pp

ed

Lin

es

/

Art

efa

cts

(7)

Acc

ep

tan

ce S

tatu

s

Comments File Name [e.g yymmdd; tbd...]






Dataset 1 (14)

LULC

Re

ad

ab

ility

(1

)

Ch

eck

He

ad

er

/Me

tad

ata

(2)

Da

ta C

ove

rage

Ma

tch

es

Se

rvic

e A

rea

(%

)

Pro

ject

ion

/ S

ph

ero

id

, E

PS

G(1

6)

Sp

ati

al R

eso

luti

on

a

nd

un

it (

e.g

. m

, k

m)

Da

ta F

orm

at

(3)

Bit

De

pth

(5)

Co

mp

lete

ne

ss

(Ve

cto

r: A

ttri

bu

te

Ta

ble

; R

ast

er:

Th

em

ati

c V

alu

es)

Ge

om

. M

isa

lign

me

nt

Dro

pp

ed

Lin

es

/ A

rte

fact

s (

EO

da

ta

on

ly)(

7)

Uti

lity

for

Cu

rre

nt

Pro

ject


EO4SD-Urban LULC Product ☒ Yes

☐ No

☒ Complete (INSPIRE/ISO19119)

☐ Incomplete

☐ not available

☒ Yes

☐ No

☐ Unknown If no: xxx%

☒ Correct

☐ Incorrect

☐ Unknown EPSG: 32749


☐Incomplete

☒ No

☐ Yes

☐ Unknown If yes: XX m

☒ No

☐ Yes

☐ n.a. If yes: xxx %

☐ None

☐ Partial

☒ Full

Comments: /






Pro

cess

ing D

ate

AO

I

Cit

y /

Re

gio

n /

Co

un

try

Pro

ject

ion

/ S

ph

ero

id (1

6)

No

. &

RM

S (

m)

of

GC

Ps

(17

)

No

. &

RM

S (

m)

of

TP

s (1

8)

No

. &

RM

S (

m)

of

CP

s (1

9)

Dig

ita

l E

leva

tio

n M

od

el

(DE

M)

Mo

de

l /

Alg

ori

thm

(13

)

Re

sam

plin

g

Me

tho

d (2

0)

Va

lida

tio

n R

ep

ort

s

Acc

ep

tan

ce S

tatu

s




1. 15DEC30080039-M2AS-056358854010_01_P002_toa_PSH

12.04.2017

Arusha TZA

UTM37S / WGS84

MGCPs 11;

RMSE x1,17; y

1,07

4; RMSE 0,61; y 0,67

N/A SRTM

30 0 order polynom

CC Yes Yes

o15DEC30080039-M2AS-056358854010_01_P002_toa_PSH

2. 16SEP17081149-M2AS-056358854010_01_P001_toa_PSH

12.04.2017

Arusha TZA

UTM37S / WGS84

MGCPs 11;

RMSE x1,17; y

1,07

4; RMSE 0,61; y 0,67

N/A SRTM 30

0 order polynom

CC Yes Yes

o16SEP17081149-M2AS-056358854010_01_P001_toa_PSH

3. 15JAN04081530-M2AS-056358854010_01_P003_toa_PSH

12.04.2017

Arusha TZA

UTM37S / WGS84

MGCPs 11;

RMSE x1,17; y

1,07

4; RMSE 0,61; y 0,67

N/A SRTM 30

0 order polynom

CC Yes Yes

o15JAN04081530-M2AS-056358854010_01_P003_toa_PSH

Quickbird-2

4. 05DEC14082246-M2AS-056358854020_01_P001_toa_PSH

12.04.2017

Arusha TZA

UTM37S / WGS84

MGCPs 15;

RMSE x1,01; y0,79

N/A N/A SRTM 30


CC Yes Yes

o05DEC14082246-M2AS-056358854020_01_P001_toa_PSH

5. 05SEP15081836-M2AS-056358854020_01_P002_toa_PSH

12.04.2017

Arusha TZA

UTM37S / WGS84

MGCPs 15;

RMSE x1,01; y0,79

N/A N/A SRTM 30


CC Yes Yes

o05SEP15081836-M2AS-056358854020_01_P002_toa_PSH




3.1.1 Data Fusion


Acc

ep

tan

ce S

tatu

s

Comments


056358854010_01_P001 16SEP17081149-P2AS-

056358854010_01_P001 16SEP17081149-M2AS-

056358854010_01_P001_toa_PSH

none




2m 0,5m 0,5m 0,5m 0,5m MS





Acc

ep

tan

ce S

tatu

s

Comments


056358854010_01_P002 15DEC30080039-P2AS-

056358854010_01_P002 15DEC30080039-M2AS-

056358854010_01_P002_toa_PSH

none









Acc

ep

tan

ce S

tatu

s Comments


056358854010_01_P003 15JAN04081529-P2AS-

056358854010_01_P003 15JAN04081530-M2AS-

056358854010_01_P003_toa_PSH none Sensor Worldview-2 Worldview-2 Worldview-2











Acc

ep

tan

ce S

tatu

s

Comments


056358854020_01_P001 05DEC14082245-P2AS-

056358854020_01_P001 05DEC14082246-M2AS-

056358854020_01_P001_toa_PSH

none









Acc

ep

tan

ce S

tatu

s

Comments


056358854020_01_P002 05SEP15081836-P2AS-

056358854020_01_P002 05SEP15081836-M2AS-

056358854020_01_P002_toa_PSH

none











3.2 Data Processing

Sensoren (8)

GE-1, WV-2, QB-2 Atmospheric Correction



OTHER Calculation:

Acc

ep

tan

ce S

tatu

s

Ba

cku

p

Comment

No.


Date

pro

cess

ed

Software / Method

pro

cess

ed

Software / Method

pro

cess

ed

Software / Method

pro

cess

ed

Software / Method

GeoEye1/Worldview-2

1. 16SEP17081149-M2AS-056358854010_01_P001


Yes No none

2. 15DEC30080039-M2AS-056358854010_01_P002


Yes No none

3. 15JAN04081530-M2AS-056358854010_01_P003


Yes No none

Quickbird-2

4. 05DEC14082246-M2AS-056358854020_01_P001


Yes No none

5. 05SEP15081836-M2AS-056358854020_01_P002


Yes No none




4.1 Classification

Sensors (8) GE-1, WV-2, QB-2

Processing Date




Acc

ep

tan

ce S

tatu

s

Ba

cku

p

Comment No.


pro

cess

ed

Software / Method

pro

cess

ed

Software / Method

pro

cess

ed

Software /

Method

pro

cess

ed

Software / Method

1.


06.10.2017 no No clouds Yes ERDAS IMAGINE

Yes


Yes


Yes Yes

2.

05DEC14082246-M2AS-056358854020_01_P001 05SEP15081836-M2AS-056358854020_01_P002

06.10.2017 no No clouds Yes ERDAS IMAGINE

Yes


Yes


Yes Yes





No. Product Name

Processing Date

Area Coverage

No

. o

f C

lass

es

Re

fere

nce

Da

ta

Sa

mp

le S

ize

Sa

mp

ling U

nit

(21

)

Sa

mp

ling

De

sign

(22

)

Sa

mp

le E

xclu

sio

n

Cri

teri

a (2

3)

Th

em

ati

c A

ccu

racy

(2

4)

Acc

ep

tan

ce S

tatu

s

Comment

1.

Urban Green Areas 2005

08.10.2017 Full AOI (urban core area)

4 VHR EO Data

215 per strata

POINT Stratified random sampling None, all used 92.56% Yes None

2.

Urban Green Areas 2015


4 VHR EO Data

215 per strata

POINT Stratified random sampling None, all used 94.88% Yes none




5.2 Error Matrices

Urban Green Area 2015 Reference Data




Totals 128 87 215

Accuracy Statistics z= 1.96 Overall Accuracy: 94.88%

95% Confidence Interval: 91.71% 98.06%

Class Name Producer’s Accuracy User’s Accuracy

0 - Non-Urban Green Area 92.19% 99.16%

1 - Urban Green Area 98.85% 89.58%

Urban Green Area 2005 Reference Data




Totals 104 111 215




0 - Non-Urban Green Area 93.27% 91.51%

1 - Urban Green Area 91.89% 93.58%




6.1 Completeness

INPUT DATA

No.






Metadata Comments

1. VHR EO Data Arusha

211 km² 100% 5 5 100% Yes none

2. Ancillary Data N/A N/A N/A N/A N/A Yes none

PRODUCTS

No.







Metadata Comments

1. Urban Green Areas 2005 and 2015 and change

Arusha 211 km² 100% 100% 0% 100% Yes none




6.2 Compliancy

Product 1 (28) Urban Green Areas

Abstract

Urban Green Areas information product contains spatial explicit information on green areas within the urban extent. Green areas include linear green features such as river alignments, hedges and trees, as well as public parks, private gardens, forested areas, etc. Urban Green Areas & Change dataset is based on Very High Resolution (VHR) satellite imagery by means of automated classification processing techniques.


Area Coverage


Country 1: Tanzania Country 1: Tanzania YES None

A) Wall-to-wall: Arusha city A) Wall-to-wall: Arusha city

Core Urban Area (211 km²) Core Urban Area (211 km²)

Area of Interest. Arusha 211 km², defined by the national user.

B) Sampling based: B) Sampling based:

N/A N/A




See EO data acquisition dates. 2005 and 2015 (+/- 1 years) 2005 and 2015/2016


No update No update



WGS84 / UTM 37S WGS84 / UTM 37S YES None



Single date

0 Non-urban green area 0 Non-urban green area YES

1 Urban green area 1 Urban green area YES

Change product

0 Non-urban green area 0 Non-urban green area YES

1 Permanent urban green area 1 Permanent urban green area YES

2 Loss of urban green area 2 Loss of urban green area YES

3 New urban green area 3 New urban green area YES

255 No data available due to absence of historic VHR imagery

255 No data available due to absence of historic VHR imagery

YES






N/A N/A

Spatial Resolution


1 m 1 m YES None



1 m² 1 m² YES None

Data Type


Raster data Raster data YES None

Bit Depth


N/A N/A

Data Format


GEOTIFF GEOTIFF YES None

Class Coding


N/A N/A

Metadata




YES None


Thematic Accuracy


> 80% > 80% YES None

Positional Accuracy


3 m (CE90) 3 m (CE90) YES None

Delivery Procedure

Service Provision


Online via FTP Online via FTP YES None

Delivery Date


End of October 2017 Third Delivery April 2018 YES None




Glossary

Quality Checks














Purpose (9)




































Lineage (29) Definition (INSPIRE 2015): Lineage is “a statement on process history and/or overall quality of the spatial data set. Where appropriate it may include a statement whether the data set has been validated or quality assured, whether it is the official version (if multiple versions exist), and whether it has legal validity. The value domain of this element is free text.”





Page 16










Service Provider: GAF Editor: A. Broszeit


Product: Planned/Unplanned Settlement Areas

Overview of QC-Sheets and Processing Steps Sheet used

Sheet filled in

Requirements






Data Quality Checks








Thematic Processing


Accuracy Assessment














0.1 Requirements

Product 1 (28) Planned and Unplanned Settlement Areas 2005 and 2015

Abstract

The product Planned and Unplanned Settlements contains spatial explicit information on the two settlement types within the core urban area. The distinction into the two settlement types is restricted to the residential area (LULC class 11), and to commercial areas (LULC class 1211) which also have a residential component. The Planned and Unplanned Settlements, and Change dataset is developed, based on Very High Resolution (VHR) satellite imagery by means of visual interpretation.


Area Coverage







Comments: None


WGS84 UTM Zone 37S


Single date

Planned settlements

Unplanned settlements

Change product

No change in planned settlement area

No change in unplanned settlement area

Expansion of planned settlement area

Expansion of unplanned settlement area

Decrease of planned settlement area

Decrease of unplanned settlement area

Unplanned to planned settlement area


n/a

Spatial Resolution






0.25 ha

Data Type & Format

Shapefile *.shp

Bit Depth

n.a.

Class Coding

Class Name RGB Code

Single date

Planned settlements 132/0/168

Unplanned settlements 168/0/0

Change product

No change in planned settlement area 132/0/168

No change in unplanned settlement area 168/0/0

Expansion of planned settlement area 197/0/255

Expansion of unplanned settlement area 255/0/0

Decrease of planned settlement area 168/132/167

Decrease of unplanned settlement area 255/100/100

Unplanned to planned settlement area 255/0/197

Metadata



Thematic Accuracy

> 80%

Positional Accuracy

RMSE < 15 m

Delivery Procedure

Service Provision

Online via FTP

Delivery Date

End of October 2017




1.1 List of EO Data


Incoming Date

Acquisition Date

Proc. Level

Path / Row





Data Format (3)

Bit Depth (5)



WV-2, GE-1

1.

16SEP17081149-M2AS_R1C1-056358854010_01_P001

31.03.2017

17.09.2016


MUL: 2m; PAN: 0.5m


.xml

16SEP17081149-M2AS_R1C2-056358854010_01_P001

31.03.2017

17.09.2016


MUL: 2m; PAN: 0.5m


.xml

16SEP17081149-M2AS_R1C3-056358854010_01_P001

31.03.2017

17.09.2016


MUL: 2m; PAN: 0.5m


.xml

16SEP17081149-M2AS_R2C1-056358854010_01_P001

31.03.2017

17.09.2016


MUL: 2m; PAN: 0.5m


.xml

16SEP17081149-M2AS_R2C2-056358854010_01_P001

31.03.2017

17.09.2016


MUL: 2m; PAN: 0.5m


.xml

16SEP17081149-M2AS_R2C3-056358854010_01_P001

31.03.2017

17.09.2016


MUL: 2m; PAN: 0.5m


.xml

2. 15DEC30080039-M2AS-056358854010_01_P002

31.03.2017

30.12.2015

LV2A Arusha, TZA

MUL: 2m; PAN: 0.5m


.xml

3. 15JAN04081530-M2AS-056358854010_01_P003

31.03.2017

04.01.2015

LV2A Arusha, TZA

MUL: 2m; PAN: 0.5m

4 4.7e-02%


.xml

Quickbird-2

4.

05DEC14082246-M2AS_R1C1-056358854020_01_P001

31.03.2017

05.12.2005


MUL: 2m; PAN: 0.6m


16 Bit

.imd,

.rpb,

.til, xml

05DEC14082246-M2AS_R1C2-056358854020_01_P001

31.03.2017

05.12.2005


MUL: 2m; PAN: 0.6m


16 Bit

.imd,

.rpb,

.til, xml

05DEC14082246-M2AS_R2C1-056358854020_01_P001

31.03.2017

05.12.2005


MUL: 2m; PAN: 0.6m


16 Bit

.imd,

.rpb,

.til, xml

05DEC14082246-M2AS_R2C2-056358854020_01_P001

31.03.2017

05.12.2005


MUL: 2m; PAN: 0.6m


16 Bit

.imd,

.rpb,

.til, xml




5.

05SEP15081836-M2AS_R1C1-056358854020_01_P002

31.03.2017

15.09.2005


MUL: 2m; PAN: 0.6m


16 Bit

.imd,

.rpb,

.til, xml

05SEP15081836-M2AS_R2C1-056358854020_01_P002

31.03.2017

15.09.2005


MUL: 2m; PAN: 0.6m


16 Bit

.imd,

.rpb,

.til, xml






Data Type (4)


No. of Sample Plots



Purpose (9)


Lineage:





Dataset 1 (14)


Metadata (2)



Area Coverag

e

Data Type (4)

Data Format (3)



No. of Classes

Thematic

Accuracy



2017-07-30 Yes


Arsuha (urban area) 100% Vector *.shp EPSG:32736, WGS 84 / UTM zone 36S

< 3 m 24 88% Yes

Lineage(29): /

Source (30):: GAF AG

Dataset 2 (14)

Report Date of Receipt from Client

Metadata (2)



Area Coverag

e

Data Type (4)

Data Format (3)



No. of Classes

Thematic

Accuracy


Arusha Master Plan 2015-2035 Volume 1: Main Report

2016-12-21 n.a. 2016-07-22 Arusha/ Tanzania Arusha Document

pdf n.a. n.a. n.a. n.a. n.a.

Lineage(29): The Master Plan of Arusha was used as a reference while mapping planned and unplanned settlements in 2015.





2.1 EO Data Quality

Sensoren(8)

WV, GE GE-1, WV-2, QB-2

Ba

cku

p

Re

ad

ab

ility

(1

)

Ch

eck

He

ad

er

/ M

eta

da

ta

(2)

Band Specifications

Pro

ject

ion

/ S

ph

ero

id (1

6)

Sce

ne

Lo

cati

on

Co

mp

lete

ne

ss o

f

Ad

dit

ion

al D

ata

Da

ta F

orm

at

(3)

Bit

De

pth

(5)

Clo

ud

Co

ver

(in

tern

al

che

ck)

Ra

dio

me

try

To

po

gra

ph

y

Dro

pp

ed

Lin

es

/ A

rte

fact

s (7

)

Acc

ep

tan

ce S

tatu

s

Comment


No

. o

f B

an

ds

Ba

nd

Re

gis

tra

tio

n

Sp

ect

ral/

Sp

ati

al

Re

solu

tio

n


1. 16SEP17081149-M2AS-056358854010_01_P001


Y - V Q

Y - V Q

Y - V

N - V Q

Yes none

2. 15DEC30080039-M2AS-056358854010_01_P002


Y - V Q

Y - V Q

Y - V

N - V Q

Yes none

3. 15JAN04081530-M2AS-056358854010_01_P003


Y - V Q

Y - V Q

Y - V

N - V Q

Yes none

Quickbird-2

4. 05DEC14082246-M2AS-056358854020_01_P001


Y - V Q

Y - V Q

Y - V

N - V Q

Yes none

5. 05SEP15081836-M2AS-056358854020_01_P002


Y - V Q

Y - V Q

Y - V

N - V Q

Yes none






Ba

cku

p

Re

ad

ab

ility

(1)

Ch

eck

He

ad

er

/ M

eta

da

ta (2

)

Ext

en

t

Pro

ject

ion

/ S

ph

ero

id

(16

)

Sp

ati

al R

eso

luti

on

Da

ta F

orm

at

(3)

Bit

De

pth

(5)

Lo

cati

on

Co

mp

lete

ne

ss

Ge

om

. M

isa

lign

me

nt

Pla

usi

bili

ty

Dro

pp

ed

Lin

es

/

Art

efa

cts

(7)

Acc

ep

tan

ce S

tatu

s







Dataset 1 (14)

LULC

Re

ad

ab

ility

(1

)

Ch

eck

He

ad

er

/Me

tad

ata

(2)

Da

ta C

ove

rage

Ma

tch

es

Se

rvic

e A

rea

(%

)

Pro

ject

ion

/ S

ph

ero

id

, E

PS

G(1

6)

Sp

ati

al R

eso

luti

on

a

nd

un

it (

e.g

. m

, k

m)

Da

ta F

orm

at

(3)

Bit

De

pth

(5)

Co

mp

lete

ne

ss

(Ve

cto

r: A

ttri

bu

te

Ta

ble

; R

ast

er:

Th

em

ati

c V

alu

es)

Ge

om

. M

isa

lign

me

nt

Dro

pp

ed

Lin

es

/ A

rte

fact

s (

EO

da

ta

on

ly)(

7)

Uti

lity

for

Cu

rre

nt

Pro

ject



☐ No


☐ Incomplete

☐ not available

☒ Yes

☐ No


☒ Correct

☐ Incorrect



☐Incomplete

☒ No

☐ Yes


☒ No

☐ Yes


☐ None

☐ Partial

☒ Full

Comments: /

Dataset 2(14) Report

Re

ad

ab

ility

(1)

Ch

eck

He

ad

er

/Me

tad

ata

(2)

Da

ta C

ove

rage

M

atc

he

s S

erv

ice

A

rea

(%

)

Pro

ject

ion

/ S

ph

ero

id

, E

PS

G(1

6)

Sp

ati

al R

eso

luti

on

a

nd

un

it (

e.g

. m

, k

m)

Da

ta F

orm

at

(3)

Bit

De

pth

(5)

Co

mp

lete

ne

ss

(Ve

cto

r: A

ttri

bu

te

Ta

ble

; R

ast

er:

Th

em

ati

c V

alu

es)

Ge

om

. M

isa

lign

me

nt

Dro

pp

ed

Lin

es

/

Art

efa

cts

(E

O d

ata

o

nly

)(7

)

Uti

lity

for

Cu

rre

nt

Pro

ject



☒ Yes

☐ No n.a. n.a. n.a. n.a. *.shp n.a. n.a. n.a. n.a.

☐ None

☐ Partial

☒ Full

Comments: /






Pro

cess

ing D

ate

AO

I

Cit

y /

Re

gio

n /

Co

un

try

Pro

ject

ion

/ S

ph

ero

id (1

6)

No

. &

RM

S (

m)

of

GC

Ps

(17

)

No

. &

RM

S (

m)

of

TP

s (1

8)

No

. &

RM

S (

m)

of

CP

s (1

9)

Dig

ita

l E

leva

tio

n M

od

el

(DE

M)

Mo

de

l /

Alg

ori

thm

(13

)

Re

sam

plin

g

Me

tho

d (2

0)

Va

lida

tio

n R

ep

ort

s

Acc

ep

tan

ce S

tatu

s




1. 15DEC30080039-M2AS-056358854010_01_P002_toa_PSH

12.04.2017

Arusha TZA

UTM37S / WGS84

MGCPs 11;

RMSE x1,17; y

1,07

4; RMSE 0,61; y 0,67

N/A SRTM

30 0 order polynom

CC Yes Yes

o15DEC30080039-M2AS-056358854010_01_P002_toa_PSH

2. 16SEP17081149-M2AS-056358854010_01_P001_toa_PSH

12.04.2017

Arusha TZA

UTM37S / WGS84

MGCPs 11;

RMSE x1,17; y

1,07

4; RMSE 0,61; y 0,67

N/A SRTM 30

0 order polynom

CC Yes Yes

o16SEP17081149-M2AS-056358854010_01_P001_toa_PSH

3. 15JAN04081530-M2AS-056358854010_01_P003_toa_PSH

12.04.2017

Arusha TZA

UTM37S / WGS84

MGCPs 11;

RMSE x1,17; y

1,07

4; RMSE 0,61; y 0,67

N/A SRTM 30

0 order polynom

CC Yes Yes

o15JAN04081530-M2AS-056358854010_01_P003_toa_PSH

Quickbird-2

4. 05DEC14082246-M2AS-056358854020_01_P001_toa_PSH

12.04.2017

Arusha TZA

UTM37S / WGS84

MGCPs 15;

RMSE x1,01; y0,79

N/A N/A SRTM 30


CC Yes Yes

o05DEC14082246-M2AS-056358854020_01_P001_toa_PSH

5. 05SEP15081836-M2AS-056358854020_01_P002_toa_PSH

12.04.2017

Arusha TZA

UTM37S / WGS84

MGCPs 15;

RMSE x1,01; y0,79

N/A N/A SRTM 30


CC Yes Yes

o05SEP15081836-M2AS-056358854020_01_P002_toa_PSH




3.1.1 Data Fusion


Acc

ep

tan

ce S

tatu

s

Comments


056358854010_01_P001 16SEP17081149-P2AS-

056358854010_01_P001 16SEP17081149-M2AS-

056358854010_01_P001_toa_PSH

none




2m 0,5m 0,5m 0,5m 0,5m MS





Acc

ep

tan

ce S

tatu

s

Comments


056358854010_01_P002 15DEC30080039-P2AS-

056358854010_01_P002 15DEC30080039-M2AS-

056358854010_01_P002_toa_PSH

none









Acc

ep

tan

ce S

tatu

s Comments


056358854010_01_P003 15JAN04081529-P2AS-

056358854010_01_P003 15JAN04081530-M2AS-

056358854010_01_P003_toa_PSH none Sensor Worldview-2 Worldview-2 Worldview-2











Acc

ep

tan

ce S

tatu

s

Comments


056358854020_01_P001 05DEC14082245-P2AS-

056358854020_01_P001 05DEC14082246-M2AS-

056358854020_01_P001_toa_PSH

none









Acc

ep

tan

ce S

tatu

s

Comments


056358854020_01_P002 05SEP15081836-P2AS-

056358854020_01_P002 05SEP15081836-M2AS-

056358854020_01_P002_toa_PSH

none











3.2 Data Processing

Sensoren (8)

GE-1, WV-2, QB-2 Atmospheric Correction



OTHER Calculation:

Acc

ep

tan

ce S

tatu

s

Ba

cku

p

Comment

No.


Date

pro

cess

ed

Software / Method

pro

cess

ed

Software / Method

pro

cess

ed

Software / Method

pro

cess

ed

Software / Method

GeoEye1/Worldview-2

1. 16SEP17081149-M2AS-056358854010_01_P001


Yes No none

2. 15DEC30080039-M2AS-056358854010_01_P002


Yes No none

3. 15JAN04081530-M2AS-056358854010_01_P003


Yes No none

Quickbird-2

4. 05DEC14082246-M2AS-056358854020_01_P001


Yes No none

5. 05SEP15081836-M2AS-056358854020_01_P002


Yes No none




4.1 Classification

Sensors (8) GE-1, WV-2, QB-2

Processing Date




Acc

ep

tan

ce S

tatu

s

Ba

cku

p

Comment No.


pro

cess

ed

Software / Method

pro

cess

ed

Software / Method

pro

cess

ed

Software /

Method

pro

cess

ed

Software / Method

1.




Yes


Yes


Yes Yes

2.

05DEC14082246-M2AS-056358854020_01_P001 05SEP15081836-M2AS-056358854020_01_P002



Yes


Yes


Yes Yes





No. Product Name

Processing Date

Area Coverage

No

. o

f C

lass

es

Re

fere

nce

Da

ta

Sa

mp

le S

ize

Sa

mp

ling U

nit

(21

)

Sa

mp

ling

De

sign

(22

)

Sa

mp

le E

xclu

sio

n

Cri

teri

a (2

3)

Th

em

ati

c A

ccu

racy

(2

4)

Acc

ep

tan

ce S

tatu

s

Comment

1. Settlements 2015


2 VHR EO Data

107 per strata

POINT Stratified random sampling None, all used 94.39% Yes




5.2 Error Matrices

Settlements 2015 Reference Data

Totals Planned Settlement Area Unplanned Settlement Area

Planned Settlement Area 97 10 107

Unplanned Settlement Area 2 105 107

Totals 99 115 214




Planned Settlement Area 97.98% 90.65%

Unplanned Settlement Area 91.30% 98.13%




6.1 Completeness

INPUT DATA

No.






Metadata Comments

1. VHR EO Data Arusha

211 km² 100% 5 5 100% Yes none

2. Ancillary Data N/A N/A N/A N/A N/A Yes none

PRODUCTS

No.







Metadata Comments

1.

Planned and Unplanned Settlement Areas 2005 and 2015 and change





6.2 Compliancy

Product 1 (28) Planned and Unplanned Settlement Areas 2005 and 2015

Abstract

The product Planned and Unplanned Settlements contains spatial explicit information on the two settlement types within the core urban area. The distinction into the two settlement types is restricted to the residential area (LULC class 11), and to commercial areas (LULC class 1211) which also have a residential component. The Planned and Unplanned Settlements, and Change dataset is developed, based on Very High Resolution (VHR) satellite imagery by means of visual interpretation.


Area Coverage







N/A N/A




See EO data acquisition dates. 2005 and 2015 (+/- 1 years) 2005 and 2015/2016


No update No update






Single date

Planned settlements Planned settlements YES

Unplanned settlements Unplanned settlements YES

Change product

No change in planned settlement area No change in planned settlement area YES

No change in unplanned settlement area No change in unplanned settlement area YES

Expansion of planned settlement area Expansion of planned settlement area YES

Expansion of unplanned settlement area Expansion of unplanned settlement area YES

Decrease of planned settlement area Decrease of planned settlement area YES

Decrease of unplanned settlement area Decrease of unplanned settlement area YES




Unplanned to planned settlement area Unplanned to planned settlement area YES



N/A N/A N/A

Spatial Resolution


N/A N/A N/A None



0.25 ha 0.25 ha YES None

Data Type


Vector data Vector data YES None

Bit Depth


N/A N/A

Data Format


ESRI SHP ESRI SHP YES None

Class Coding


N/A N/A

Metadata




YES None


Thematic Accuracy


> 80% > 80% YES None

Positional Accuracy


RMSE < 15 m 15m YES None

Delivery Procedure

Service Provision



Delivery Date





End of October 2017 End of October 2017 YES None




Glossary

Quality Checks














Purpose (9)














































Service Provider: GAF Editor: D. Angelova


Product: Population Distribution and Density

Overview of QC-Sheets and Processing Steps Sheet used

Sheet filled in

Requirements






Data Quality Checks








Thematic Processing


Accuracy Assessment














0.1 Requirements

Product 1 (28) Population Distribution and Density Change 2005 and 2015

Abstract

The Population Distribution and Density Change product contains spatial explicit information about population distribution changes at a certain degree, within the Core Urban Districts of Arusha. The following eight change classes are identified, based on a frequency distribution analysis of the changes that occur: Unchanged Population Distribution; Up to –100% decrease; Up to 200% increase; 201% - 400% increase; 401% - 600% increase; 601% - 800% increase; 801% - 1000% increase; More than 1000% increase. The Population Distribution and Density Change product is estimated for all objects which belong to the residential class (LULC class 11), using 5 different data sources: LULC Product, WorldPop data (2015) with a spatial resolution of 100m, official population census data from 2002 and 2012, Soil Sealing layer and Administrative Boundaries.


Area Coverage



Area km² Core Urban: 211 km2 B) Sampling based: n/a




Comments: None


WGS84 UTM Zone 37S


Single date

1 Population Density between 0 – 150 inhabitants/ km2








9 Population Density between 18001 - 41000 inhabitants/ km2

Change product

1 Unchanged Population Distribution

2 Up to -100% decrease

3 Up to 200% increase

4 201% - 400% increase




5 401% - 600% increase

6 601% - 800% increase

7 801% - 1000% increase

8 More than 1000% increase


n/a

Spatial Resolution



n.a.

Data Type & Format

Shapefile *.shp

Bit Depth

n.a.

Class Coding


Single date

1 Population Density between 0 – 150 inhabitants/ km2 242/241/162








9 Population Density between 18001 - 41000 inhabitants/ km2 93/26/201

Change product

1 Unchanged Population Distribution 156/156/156

2 Up to -100% decrease 112/153/89

3 Up to 200% increase 242/237/194

4 201% - 400% increase 237/215/114

5 401% - 600% increase 242/176/99

6 601% - 800% increase 237/175/175

7 801% - 1000% increase 196/113/96

8 More than 1000% increase 120/35/35

Metadata



Thematic Accuracy

> 80%

Positional Accuracy

RMSE < 15 m




Delivery Procedure

Service Provision

Online via FTP

Delivery Date

End of October 2017




1.1 List of EO Data

Sensoren (8) No Data used

Incoming Date

Acquisition Date

Proc. Level

Path / Row


Spatial Res.

No. of Bands



Data Format (3)

Bit Depth (5)



No EO Data Used






Data Type (4)


No. of Sample Plots



Purpose (9)


Lineage:





Dataset 1 (14)


Metadata (2)



Area Coverage

Data Type (4)

Data Format (3)



No. of Classes

Thematic Accuracy



2017-07-30 Yes


Arusha (urban area) 100% Vector *.shp EPSG:32736, WGS 84 / UTM zone 37S

< 3 m 24 88% Yes

Lineage(29): The Urban Land Use/Land Cover Baseline Product and the derived Soil Sealing layer are used to interpolate the population distribution and density for each residential unit, within the Core Urban Districts of Arusha.

Source (30): GAF AG

Dataset 2 (14)


Metadata (2)



Area Coverage

Data Type (4)

Data Format (3)



No. of Classes

Thematic Accuracy



2016-12-21 n.a. 2016-07-22 Arusha/ Tanzania Arusha Document


Lineage(29): The Arusha Master Plan 2015-2035 is used to define class thresholds for both current and historic status of the population density in Arusha.

Source (30): Received from MaryGrace Weber.

Dataset 3 (14)

Population Raster

Dataset


Metadata (2)



Area Coverage

Data Type (4)

Data Format (3)



No. of Classes

Thematic Accuracy


WorldPop Dataset 2017-09-01 Yes 2015 Arusha/Tanzania 100% Raster Geotiff Geographic, WGS84

n.a. n.a. n.a. n.a.

Lineage(29): The WorldPop Dataset is used to interpolate the current population data for each residential object in the Core Urban Districts of Arusha.

Source (30): Downloaded from WorldPop: http://www.worldpop.org.uk/




Dataset 4 (14)


Metadata (2)



Area Coverage

Data Type (4)

Data Format (3)



No. of Classes

Thematic Accuracy


2012 POPULATION AND HOUSING CENSUS

2017-09-01 n.a. 2012-08-26 Tanzania National coverage

Document


Lineage(29): The 2012 Population and Housing Census report is used to project the number of inhabitants per district, and to adjust the population residuals in the current Population Distribution and Density product.

Source (30): Downloaded from United Republic of Tanzania, Ministry of East African Cooperation: http://meac.go.tz/

Dataset 5 (14)


Metadata (2)



Area Coverage

Data Type (4)

Data Format (3)



No. of Classes

Thematic Accuracy



2017-09-01 Yes

(The World Bank)

2002-08-22 Tanzania National coverage

Document


Lineage(29): The 2002 Population and Housing Census report is used to interpolate the population data in the historic Population Distribution and Density product.

Source (30): Downloaded from the IPUMS – International Website: https://international.ipums.org/international/

Dataset 6 (14)

Administrative

Boundaries


Metadata (2)



Area Coverage

Data Type (4)

Data Format (3)



No. of Classes

Thematic Accuracy


TZA_adm3 2017-09-01 Yes

2015 Tanzania

National coverage

Vector *.shp WGS 84 unknown n.a. unknown Yes

Lineage(29): This shapefile provides with information, regarding the administrative boundaries at a district level in Tanzania.

Source (30): Downloaded from Global Administrative Areas: http://www.gadm.org/country




2.1 EO Data Quality

Sensoren(8)

WV, GE No Data used

Ba

cku

p

Re

ad

ab

ility

(1

)

Ch

eck

He

ad

er

/ M

eta

da

ta

(2)

Band Specifications

Pro

ject

ion

/ S

ph

ero

id (1

6)

Sce

ne

Lo

cati

on

Co

mp

lete

ne

ss o

f A

dd

itio

na

l D

ata

Da

ta F

orm

at

(3)

Bit

De

pth

(5)

Clo

ud

Co

ver

(in

tern

al

che

ck)

Ra

dio

me

try

To

po

gra

ph

y

Dro

pp

ed

Lin

es

/ A

rte

fact

s (7

)

Acc

ep

tan

ce S

tatu

s

Comment


No

. o

f B

an

ds

Ba

nd

Re

gis

tra

tio

n

Sp

ect

ral/

Sp

ati

al

Re

solu

tio

n

No EO Data Used

n.a n.a n.a n.a n.a n.a n.a n.a n.a n.a n.a n.a n.a n.a n.a n.a n.a n.a






Ba

cku

p

Re

ad

ab

ility

(1)

Ch

eck

He

ad

er

/ M

eta

da

ta (2

)

Ext

en

t

Pro

ject

ion

/ S

ph

ero

id

(16

)

Sp

ati

al R

eso

luti

on

Da

ta F

orm

at

(3)

Bit

De

pth

(5)

Lo

cati

on

Co

mp

lete

ne

ss

Ge

om

. M

isa

lign

me

nt

Pla

usi

bili

ty

Dro

pp

ed

Lin

es

/

Art

efa

cts

(7)

Acc

ep

tan

ce S

tatu

s


No In-situ Data Used n.a n.a n.a n.a n.a n.a n.a n.a n.a n.a n.a n.a n.a n.a /





Dataset 1 (14) LULC

Re

ad

ab

ility

(1

)

Ch

eck

He

ad

er

/Me

tad

ata

(2)

Da

ta C

ove

rage

M

atc

he

s S

erv

ice

A

rea

(%

)

Pro

ject

ion

/ S

ph

ero

id

, E

PS

G(1

6)

Sp

ati

al R

eso

luti

on

a

nd

un

it (

e.g

. m

, k

m)

Da

ta F

orm

at

(3)

Bit

De

pth

(5)

Co

mp

lete

ne

ss

(Ve

cto

r: A

ttri

bu

te

Ta

ble

; R

ast

er:

Th

em

ati

c V

alu

es)

Ge

om

. M

isa

lign

me

nt

Dro

pp

ed

Lin

es

/ A

rte

fact

s (

EO

da

ta

on

ly)(

7)

Uti

lity

for

Cu

rre

nt

Pro

ject



☐ No


☐ Incomplete

☐ not available

☒ Yes

☐ No


☒ Correct

☐ Incorrect



☐Incomplete

☒ No

☐ Yes


☒ No

☐ Yes


☐ None

☐ Partial

☒ Full

Comments:

Dataset 2 (14) Report

Re

ad

ab

ility

(1)

Ch

eck

He

ad

er

/Me

tad

ata

(2)

Da

ta C

ove

rage

M

atc

he

s S

erv

ice

A

rea

(%

)

Pro

ject

ion

/ S

ph

ero

id

, E

PS

G(1

6)

Sp

ati

al R

eso

luti

on

an

d u

nit

(e

.g.

m, k

m)

Da

ta F

orm

at

(3)

Bit

De

pth

(5)

Co

mp

lete

ne

ss

(Ve

cto

r: A

ttri

bu

te

Ta

ble

; R

ast

er:

Th

em

ati

c V

alu

es)

Ge

om

. M

isa

lign

me

nt

Dro

pp

ed

Lin

es

/ A

rte

fact

s (

EO

da

ta

on

ly)(

7)

Uti

lity

for

Cu

rre

nt

Pro

ject



☒ Yes

☐ No

☐ Complete

☐ Incomplete

☒ not available

☒ Yes

☐ No


n.a. n.a. *.pdf n.a. n.a. n.a. n.a.

☐ None

☒ Partial

☐ Full

Comments: The Master Plan is used to define proper thresholds for the population density classes in Arusha.




Dataset 3 (14) Population Raster Dataset

Re

ad

ab

ility

(1)

Ch

eck

He

ad

er

/Me

tad

ata

(2)

Da

ta C

ove

rage

M

atc

he

s S

erv

ice

Are

a

(%)

Pro

ject

ion

/ S

ph

ero

id

, E

PS

G(1

6)

Sp

ati

al R

eso

luti

on

an

d u

nit

(e

.g.

m, k

m)

Da

ta F

orm

at

(3)

Bit

De

pth

(5)

Co

mp

lete

ne

ss

(Ve

cto

r: A

ttri

bu

te

Ta

ble

; R

ast

er:

Th

em

ati

c V

alu

es)

Ge

om

. M

isa

lign

me

nt

Dro

pp

ed

Lin

es

/ A

rte

fact

s (

EO

da

ta

on

ly)(

7)

Uti

lity

for

Cu

rre

nt

Pro

ject


WorldPop Dataset ☒ Yes

☐ No

☐ Complete

☒ Incomplete

☐ not available

☒ Yes

☐ No


☒ Correct

☐ Incorrect


100m *.tif 32 Bit ☒ Complete

☐Incomplete

☒ No

☐ Yes


☒ No

☐ Yes


☐ None

☒ Partial

☐ Full

Comments: The WorldPop Dataset is used only for the estimation of the current Population Distribution and Density product.


Re

ad

ab

ility

(1

)

Ch

eck

He

ad

er

/Me

tad

ata

(2)

Da

ta C

ove

rage

M

atc

he

s S

erv

ice

Are

a (

%)

Pro

ject

ion

/ S

ph

ero

id

, E

PS

G(1

6)

Sp

ati

al R

eso

luti

on

an

d u

nit

(e

.g.

m, k

m)

Da

ta F

orm

at

(3)

Bit

De

pth

(5)

Co

mp

lete

ne

ss

(Ve

cto

r: A

ttri

bu

te

Ta

ble

; R

ast

er:

Th

em

ati

c V

alu

es)

Ge

om

. M

isa

lign

me

nt

Dro

pp

ed

Lin

es

/

Art

efa

cts

(E

O d

ata

o

nly

)(7

)

Uti

lity

for

Cu

rre

nt

Pro

ject



☒ Yes

☐ No

☐ Complete

☐ Incomplete

☒ not available

n.a. n.a. n.a. *.pdf n.a. n.a. n.a. n.a.

☐ None

☐ Partial

☒ Full

Comments: The 2012 Population Census report is used to derive the number of inhabitants per district in 2015 in Arusha.


Re

ad

ab

ility

(1)

Ch

eck

He

ad

er

/Me

tad

ata

(2)

Da

ta C

ove

rage

M

atc

he

s S

erv

ice

Are

a (

%)

Pro

ject

ion

/ S

ph

ero

id

, E

PS

G(1

6)

Sp

ati

al R

eso

luti

on

an

d u

nit

(e

.g.

m, k

m)

Da

ta F

orm

at

(3)

Bit

De

pth

(5)

Co

mp

lete

ne

ss

(Ve

cto

r: A

ttri

bu

te

Ta

ble

; R

ast

er:

Th

em

ati

c V

alu

es)

Ge

om

. M

isa

lign

me

nt

Dro

pp

ed

Lin

es

/

Art

efa

cts

(E

O d

ata

o

nly

)(7

)

Uti

lity

for

Cu

rre

nt

Pro

ject



☒ Yes

☐ No

☒ Complete

☐ Incomplete

☐ not available

n.a. n.a. n.a. *.pdf n.a. n.a. n.a. n.a.

☐ None

☐ Partial

☒ Full

Comments: The 2002 Population Census report is used to derive the number of inhabitants per district in 2005 in Arusha.




Dataset 6 (14)

Administrative Boundaries

Re

ad

ab

ility

(1)

Ch

eck

He

ad

er

/Me

tad

ata

(2

)

Da

ta C

ove

rage

Ma

tch

es

Se

rvic

e A

rea

(%

)

Pro

ject

ion

/ S

ph

ero

id ,

E

PS

G(1

6)

Sp

ati

al R

eso

luti

on

an

d u

nit

(e.g

. m

, k

m)

Da

ta F

orm

at

(3)

Bit

De

pth

(5)

Co

mp

lete

ne

ss (

Ve

cto

r:

Att

rib

ute

Ta

ble

; R

ast

er:

Th

em

ati

c V

alu

es)

Ge

om

. M

isa

lign

me

nt

Dro

pp

ed

Lin

es

/ A

rte

fact

s

(EO

da

ta o

nly

)(7

)

Uti

lity

for

Cu

rre

nt

Pro

ject


TZA_adm3 ☒ Yes

☐ No

☐ Complete

☒ Incomplete

☐ not available

☒ Yes

☐ No


☒ Correct

☐ Incorrect



☐Incomplete

☒ No

☐ Yes


☒ No

☐ Yes


☐ None

☐ Partial

☒ Full

Comments: /




3.1.1 Data Fusion


Acc

ep

tan

ce S

tatu

s

Comments

Filename n.a n.a n.a

none

Sensor n.a n.a n.a

Method n.a n.a n.a

Spatial Resolution / (MS/Pan) n.a n.a n.a n.a n.a n.a

Band Combination n.a n.a n.a

Data Format (3) & Bit Depth (5) n.a n.a n.a n.a n.a n.a n.a


Sensoren (8) No Data used

Pro

cess

ing D

ate

AO

I

Cit

y /

Re

gio

n /

Co

un

try

Pro

ject

ion

/ S

ph

ero

id (1

6)

No

. &

RM

S (

m)

of

GC

Ps

(17

)

No

. &

RM

S (

m)

of

TP

s (1

8)

No

. &

RM

S (

m)

of

CP

s (1

9)

Dig

ita

l E

leva

tio

n M

od

el

(DE

M)

Mo

de

l /

Alg

ori

thm

(13

)

Re

sam

plin

g

Me

tho

d (2

0)

Va

lida

tio

n R

ep

ort

s

Acc

ep

tan

ce S

tatu

s

Output File Name [e.g yymmdd; tbd...] No. File Name [e.g yymmdd; tbd...]

No EO Data Used




3.2 Data Processing

Sensoren (8)

No Data used Atmospheric Correction



OTHER Calculation:

Acc

ep

tan

ce S

tatu

s

Ba

cku

p

Comment

No. File Name [e.g yymmdd; tbd...] Processing

Date

pro

cess

ed

Software / Method

pro

cess

ed

Software / Method

pro

cess

ed

Software / Method

pro

cess

ed

Software / Method

No EO Data Used

4.1 Classification

Sensors (8) No Data used

Processing Date




Acc

ep

tan

ce S

tatu

s

Ba

cku

p

Comment

No. File Name [e.g yymmdd; tbd...]

pro

cess

ed

Software / Method

pro

cess

ed

Software / Method

pro

cess

ed

Software / Method

pro

cess

ed

Software / Method

No EO Data Used


No. Product Name Processing Date

Area Coverage

No

. o

f C

lass

es

Re

fere

nce

Da

ta

Sa

mp

le S

ize

Sa

mp

ling U

nit

(21

)

Sa

mp

ling

De

sign

(22

)

Sa

mp

le E

xclu

sio

n

Cri

teri

a (2

3)

Th

em

ati

c A

ccu

racy

(2

4)

Acc

ep

tan

ce S

tatu

s

Comment

1.

Population Distribution and Density Change 2005 and 2015

13.10.2017

Full AOI (Core Urban Districts)

8 n.a n.a n.a n.a n.a n.a Yes




5.2 Error Matrices

The estimated population per district corresponds to the 2012 population and housing census data.

In a disaggregation procedure, such as the one herein described, the outcome of the process is never less accurate than the original source data. By disaggregating

numerical data from one coarse geometry to a finer geometry, we always gain detail and approximate ground truth without the risk of deteriorating the source

information. The degree to which the disaggregation approximates reality, however, varies greatly, and it depends chiefly on: 1) the quality of the ancillary data

and 2) the appropriateness of the disaggregation algorithm and its parameters.




6.1 Completeness

INPUT DATA

No. Item AOI Coverage [km²]





Metadata Comments

1. Ancillary Data Arusha 211 km² 100% 1 1 100% Yes none

PRODUCTS

No. Product (28) AOI Coverage [km²]






Metadata Comments

1.

Population Distribution and Density Change 2005 and 2015





6.2 Compliancy

Product 1 (28) Population Distribution and Density Change 2005 and 2015

Abstract

The Population Distribution and Density Change product contains spatial explicit information about population distribution changes at a certain degree, within the Core Urban Districts of Arusha. The following eight change classes are identified, based on a frequency distribution analysis of the changes that occur: Unchanged Population Distribution; Up to –100% decrease; Up to 200% increase; 201% - 400% increase; 401% - 600% increase; 601% - 800% increase; 801% - 1000% increase; More than 1000% increase. The Population Distribution and Density Change product is estimated for all objects which belong to the residential class (LULC class 11), using 5 different data sources: LULC Product, WorldPop data (2015) with a spatial resolution of 100m, official population census data from 2002 and 2012, Soil Sealing layer and Administrative Boundaries.


Area Coverage







N/A N/A



Baseline Year(s): Baseline Year(s): Yes None

2005 and 2015 (+/- 1 years) 2005 and 2015

B) Update Frequency B) Update Frequency Yes None

No update No update






Single date

Population Density between 0 – 150 inhabitants/ km2 0 – 150 inhabitants/ km2 YES








Population Density between 18001 - 41000 inhabitants/ km2 18001 - 41000 inhabitants/ km2 YES




Change product

Unchanged Population Distribution Unchanged Population Distribution YES

Up to -100% decrease Up to -100% decrease YES

Up to 200% increase Up to 200% increase YES

201% - 400% increase 201% - 400% increase YES




More than 1000% increase More than 1000% increase YES



N/A N/A N/A

Spatial Resolution


N/A N/A N/A None



N/A N/A N/A None

Data Type


Vector data Vector data YES None

Bit Depth


N/A N/A

Data Format


ESRI SHP ESRI SHP YES None

Class Coding


N/A N/A

Metadata




YES None


Thematic Accuracy


> 80% The estimated population per district corresponds to the 2012 population and housing census data.

YES None

Positional Accuracy





RMSE < 15 m 15m YES None

Delivery Procedure

Service Provision



Delivery Date


End of October 2017 End of October 2017 YES None




Glossary

Quality Checks














Purpose (9)






































Date post:	09-Jul-2020
Category:	Documents
Upload:	others
View:	2 times
Download:	0 times