Drones for River AssessmentProfessor Ian Maddock and Sophie Pearce

River Science Research GroupSchool of Science and the Environment, University of Worcester

Outline1. Introduction: Who we are and what we work on

2. Traditional mapping and monitoring methods

3. Drones for water velocity measurement: Sophie Pearce

4. Drones for surveying topography and river morphology

5. Drones for monitoring channel change

6. Final thoughts

7. Discussion

Research Focus:

Hydromorphology & Physical HabitatThe interaction of hydrology and channel morphology and its

influence on aquatic biota (esp. macroinvertebrates & fish)

e.g. the flow

regime, drought

& floods, depth

& velocity

e.g. channel

shape, river bed

morphology,

substrate size

e.g. aquatic species,

macroinvertebrates,

plants, fish, etc.

How can we measure, survey & quantify hydraulics, morphology & physical habitat?

San Pedro River, Chile: The influence of channel marginal

habitats on native fish at a planned hydropower site

Lead Collaborator: Prof Evelyn Habit Funders: University of Concepcion, FONDECYT, Chile; UW (studentship)

Publications: Wilkes, Maddock, Link & Habit (2016), Woodget, Austrums, Maddock & Habit (2017)

Endangered Native Fish

Galaxiids

Megan Klaar, PhD (2010). Learning from unimpacted sites: Glacier

Bay, Alaska: The relationship between stream age, channel

morphology, trees in the river (large wood), native fish

Funders: UW & Uni Birmingham (PhD studentship), Prof Sandy Milner (UB)

Publications: Klaar, Maddock & Milner (2009), Klaar, Hill, Maddock & Milner (2011)

Setting environmental flows to

improve habitat availability for

fish (trout and grayling) and four

species of macroinvertebrates

Image:

https://twitter.com/envagencymids/status/966339060766711808

Upper River Derwent, Derbyshire: Assessing the impact of dam

releases on fish and macroinvertebrates

Lead Collaborators: Rachel Spence, Tim Pickering (EA), Dr Melanie Bickerton (UB)Funders: Environment Agency

Publication: Maddock, Bickerton, Spence & Pickering (2001)

Chris Greensmith (MRes in River Science): New Project;

Environmental DNA (eDNA) for detection of otter, water vole and mink

With the kind support of a BHS, JBA & EA Masters studentship award

The need for Physical Habitat Assessment?

Environmental Flows

River Restoration

Weir Removal

What do traditional surveys entail?• 1-10s kms

• Walking the river bank

or in-channel

• Mapping ‘key’ habitat

features

Pool

Pool

Pool

Glide

RiffleRun

• Using established

survey / classification

systems

Surveying, mapping and measuring rivers and their habitats at two scales:

1. Long reach habitat (‘walkover’) survey

2: Traditional mapping and monitoring methods

1. Long reach habitat (‘walkover’) survey

Scale: 1 – >10kms

Surveying, mapping and measuring rivers and their habitats at two scales:

2. Short reach hydraulic & river habitat modelling

• Reach lengths typically 50-500m

• Hundreds of measurements of topography,

velocity and depth across cross-sections

Most surveying methods:

Time consuming

Sparse point data

Low resolution of topography / bathymetry data

You need access to all parts of the site

▪ 2010 purchased our first drone

▪ 2011 CAA permission for drone flying, four ‘licenced’ pilots, with one

more next year

▪ Used in the UK and overseas: Chile, Slovenia & Switzerland

▪ Two international conferences (2013 and 2016)

▪ Three different rotary-winged, one fixed-wing UAV

Drones at the University of Worcester (UW)

Most common commercial drone usage:

Oblique aerial photos for a site overview

Most common commercial drone usage:

Video for marketing (River Teme)

Drones and Surface Flow Velocity measurement?

Using drone video for image velocimetry

3: Drones for velocity measurement

Background• Rivers need to be monitored during flood to inform prediction,

management and mitigation

• Flood conditions often pose a threat to traditional monitoring

instruments or people.

• Need to be measured for both Discharge readings and point

measurements (mainly for applications of flow to other studies,

such as calibrating hydraulic models for habitat studies)

Problems…- Time consuming with current meters

- High flows…ADCPs struggle…Gauging stations unreliable

Photo courtesy of Nick Everard

(EA)

Unreliable at high flow…• Flood warning systems rely on accurate hydrographs

• Hydrographs often based on measurements at ‘normal’ flow

conditions, and extrapolated to represent extreme flow

conditions = high errors for prediction, mitigation and

management.

Jodeau et al (2017)

Using remote technologies for river

monitoring

- Image-VelocimetryNon contact & safe, quick, direct, but errors remain

Lots of different algorithms but theoretically

very similar concepts

Three main ones to know are:

Large Scale Particle Image Velocimetry (LSPIV)

Space Time Image Velocimetry (STIV)

Particle Tracking Velocimetry (PTV)

Image-Velocimetry 1. Video recorded of the flow

(with surface features)

2. Image rectification

3. Image processing

– Calculation of

surface

velocity

Image-Velocimetry…requires surface

features

River Bow,

Canada

Bow Falls,

Canada

River Arrow,

Warwickshire

Broom monitoring

Station

Application…

Instantaneous surface velocity results

Average surface velocity results

Research and Development…

Currently lots of error sources regarding…

a) Visualisation of the water surface – features required…seeding

densities?

b) Which is more accurate drone or bankside camera?

c) Towards a completely remote technique – currently, ground truthing

still required

d) To allow a simple and effective technique…currently predominantly

restricted to expert research use…tools are not necessarily user

friendly.

Potential…

• Currently image-velocimetry is being

researched and adopted into hydrological

practises in many other countries including

Japan, France, Australia and the USA.

• EA very interested in the use of image-

velocimetry for gauging high flow events

• Has the potential to provide us with an

accurate, quick and safe method for

collecting data during events whereby access

to the river would be too dangerous….

• As well as exploring its application for during

a range of river flows (including low

flows…applications for calibrating hydraulic

models for example.

• Create topographic

maps

• Survey river beds

• Repeat flights: compare

surveys, measure and

map channel change

Drones and Structure-from-Motion (SfM) photogrammetry:

take measurements from photographs

How does it work?

4: Drones for topography and river morphology

Process the photos using

photogrammetry software:

Agisoft Photoscan

• Stitch them together.

• Remove distortions.

• Create a point cloud. Millions

of survey points…

River Teme: Point cloud >55 million survey points

River Teme: Orthophoto – high resolution, no distortion

River Teme: Orthophoto

– usable for long reach habitat mapping

Can’t we just use Google Maps / Earth satellite image?

Can’t we just use Google Maps / Earth satellite image?

Can’t we just use Google Maps / Earth satellite image?

• Image resolution is too coarse

Date of imagery is fixed.

• Drones give us bespoke flights

• Present day data on habitat

features, morphology etc

• Repeat flights for monitoring

change

River Teme: Digital Elevation Model (DEM):create cross profiles, long profile, export to hydraulic and habitat models

Any project where you need ultra-high resolution topographich survey

Detailed bathymetry maps if the conditions are suitable

Woodget

(2015)

Technological Developments for Monitoring Channel Change

• Understand natural river processes

• Identify sources of erosion and deposition

• Excessive sediment inputs leads to ecological decline

5: Drones for geomorphic change detection

Using repeat drone surveys to monitor channel migration

Geomorphic Change Detection

Before and after surveys

Map and measure erosion,

deposition, channel change,

bank erosion

(Tamminga et al 2015)

Potential for assessing

changes in morphology (and

velocities) around weir

removals

Megan Robertson (MRes in River Science): New Project; Natural Flood

Management and assessing sediment delivery to rivers using drones

Funder: Environment Agency

Final thoughts…

• These are important new tools in the river scientists toolbox

• Ultra High Resolution Data: Understand geomorphic and

hydraulic processes, map and monitor habitats and measure

flow velocity in ways not possible previously

New technology has and is transforming how we measure, map and

monitor rivers including the depth of shallow sites, velocities, habitats and

channel change

• Drone video for surface velocity measurement: Image Velocimetry

• Drone photos for topographic surveys: SfM Photogrammetry

• Repeat surveys for geomorphic change detection and soil erosion

Thank you