8/8/2019 15arspc Submission 19

1/12

8/8/2019 15arspc Submission 19

2/12

8/8/2019 15arspc Submission 19

3/12

8/8/2019 15arspc Submission 19

4/12

4

Figure 2. Photo of the moss bed in the Antarctic Special Protected Area (ASPA) 135 near Casey.

Figure 3. Photograph of a small depression highlighting the effect of micro-topography and water availability on bryophyte health.

8/8/2019 15arspc Submission 19

5/12

5

Unmanned Aerial Vehicle (UAV)

We found that the spatial scale of the moss beds (tens of m 2) makes satelliteimagery (even very high resolution QuickBird imagery with a 0.6 m pixel size)unsuitable for mapping their extent and health in sufficient detail. We thereforedeveloped a new UAV consisting of a small electric remote controlled (RC)helicopter capable of carrying three different sensors. The RC helicopter is anelectric hobby model helicopter, Align Trex 500, capable of lifting 1.5 kg with aflight time between 6 and 10 minutes (Figure 4 ). The UAV is manually flown,however, the flight is stabilised by a Helicommand 3A attitude stabilisationsystem 1. The UAV flight path is recorded by a small SkyTraq USB GPS dongle 2 (18 grams) logging at 1Hz. The flight path can be downloaded after the flightand visualised in Google Earth or GIS software (Figure 5 ). The aluminiumcamera mount was custom-built to hold three separate cameras. The UAV cancarry one camera at a time. We tested a variety of vibration isolation devices to

reduce motion blur in the photographs. Bungy cord was found to give asatisfactory and efficient result.

Figure 4. Our UAV at the study site mounted with a camera, GPS receiver, and flight stabiliser.

The camera system consists of a Canon Powershot G10 camera, capturing 15megapixel photographs every ~3 seconds. The G10 is a high-end compactcamera weighing 355 grams. Photos were triggered with an URBI USBcontroller 3 connected to the G10 and the receiver on the UAV. A switch on the

1 http://en.helicommand.com 2

http://www.blip.com.au/item.aspx?itemid=62&ref=13 3 http://www.blip.com.au/item.aspx?itemid=14&ref=8

http://en.helicommand.com/http://en.helicommand.com/http://en.helicommand.com/http://www.blip.com.au/item.aspx?itemid=62&ref=13http://www.blip.com.au/item.aspx?itemid=62&ref=13http://www.blip.com.au/item.aspx?itemid=62&ref=13http://www.blip.com.au/item.aspx?itemid=14&ref=8http://www.blip.com.au/item.aspx?itemid=14&ref=8http://www.blip.com.au/item.aspx?itemid=14&ref=8http://www.blip.com.au/item.aspx?itemid=14&ref=8http://www.blip.com.au/item.aspx?itemid=62&ref=13http://en.helicommand.com/

8/8/2019 15arspc Submission 19

6/12

6

transmitter allowed for remote and continuous shutter release. The advantageof the G10 is that it provides high quality photographs with low noise (comparedto other compact cameras) and it has many options for manual adjustments.We set the camera to aperture priority using the widest aperture and an ISO

value between 200 and 800 to capture photographs at a shutter speed of1/1000 or faster in order to reduce motion blur. At a flying height of 50 m the 28mm focal length lens captured aerial photographs with a spatial extent of 64 mby 33 m at 1.5 cm resolution. The time on the camera was synchronised withGPS time so that each photo could be geotagged afterwards. In addition, smallorange markers were used as ground control point (GCPs) for accurategeoregistration of the UAV photographs (Figure 6 ).

Figure 5. UAV flight path in Google Earth.

A second G10 camera was converted to a UV+VIS+NIR camera 4 by removal ofthe infrared filter in front of the sensor. With the addition of an XNite-BPG58 5 band pass filter (spectral range: 710 855 nm with a peak at 820 nm) we wereable to obtain NIR photographs (Figure 7 ). The combination of true colour andNIR photography allows for generation of false colour composites and basicvegetation indices such as NDVI (Nagai et al. 2009).The third sensor is a thermal camera, the FLIR Photon 320, usingmicrobolometer technology. It generates 320x240 resolution thermal videoframes at 9 Hz. Berni et al. (2009) successfully used thermal imagery acquiredfrom a UAV to assess crop health and detect water stress over agriculturalareas. We aim to use thermal imagery to assess and map the heat regulation ofbryophytes, reasoning that healthy mosses are warmer than their surroundingenvironment on cold sunny days. The thermal imagery was calibrated with

4

http://www.maxmax.com/ 5 http://www.maxmax.com/BP_Color_Series.htm

http://www.maxmax.com/http://www.maxmax.com/http://www.maxmax.com/http://www.maxmax.com/BP_Color_Series.htmhttp://www.maxmax.com/BP_Color_Series.htmhttp://www.maxmax.com/BP_Color_Series.htmhttp://www.maxmax.com/BP_Color_Series.htmhttp://www.maxmax.com/

8/8/2019 15arspc Submission 19

7/12

7

temperature measurements on the ground using a thermal radiometer and acontact k-probe thermometer.

Figure 6. Georeferenced UAV photograph of the ASP135 moss bed.

Figure 7. Two detailed UAV photographs of 6 m by 4 m areas showing a visible photograph on the left and a near-infrared photograph on the right. The NIR

photograph shows healthy moss as bright pixels.

Sampling

The influence of terrain (slope, solar radiation, hydrological conditions) onbryophyte health is very important. Differences in these terrain characteristicsmight explain the variability between the moss samples within a site andbetween sites. To determine these local environmental terrain characteristics avery high resolution digital elevation model (DEM) is required. In the 2010

8/8/2019 15arspc Submission 19

8/12

8

season we collected terrain data using highly accurate geodetic GPSequipment. Two Leica GPS1200 receivers were used in dual frequency real-time kinematic mode, providing centimetre positional and height accuracies.Static GPS points were recorded for small aluminium markers that were

installed during previous field seasons. These markers indicate the location of20x20 cm moss quadrats that have been photographed in the past and areused for photo monitoring and change detection. Accurately locating thesemoss quadrats will allow us to compare their locations to terrain layers and UAVphotography in the future. In addition, 40 bright orange aluminium disks with adiameter of 10 cm were spread out over the study area. These disks act asground control points with which the UAV aerial photographs weregeoreferenced (Figure 6 ).

To capture sufficient data for terrain interpolation we collected ~10,000 GPSpoints by walking transects logging points at 1Hz. The point density on the

transects was 0.5 m and between transects 1 5 m (radiating out from themoss bed centre, see Figure 8 a).

Digital Elevation Model

A digital elevation model (DEM) at 0.5 m resolution was interpolated from GPSheights using the Topo to Raster interpolation tool in ArcGIS resulting in ahydrological sound terrain surface. This technique is based on the ANUDEMinterpolation algorithm (Hutchinson 1989). From the DEM important topographicinformation such as aspect, slope, solar radiation, wetness index, and othermorphometric characteristics can be computed for each grid cell. Slope refers

to the rate of change in the elevation, while aspect identifies the direction of thatchange. Slope and aspect are both parameters that contribute to thecomputation of solar radiation and surface wetness. We hypothesize that thespatial patterns of moss growth are mainly driven by water availability(represented by the wetness index) and temperature (represented by the solarradiation). The resulting 50 cm resolution DEM is shown in Figure 8 b and c.

Wetness Index

We calculated the topographic wetness index (WI) to provide an indication ofthe spatial distribution of potential surface wetness caused by snow melt. Thewetness index is calculated from a local drainage direction (LDD) network andthe slope layer, both derived from a DEM. The LDD network shows the flowdirection for each grid cell and based on this layer the upstream area for eachcell can be calculated. The wetness index is a function of the upstream areaand the slope, whereby the greater the upstream area and the flatter the cell thehigher the wetness index. The wetness index only provides a hypotheticalsurface wetness (unitless index) without taking into account soil properties orgeological characteristics, however, it is a very useful DEM derivativehighlighting the relative water distribution with topographic features as the keydriving force. The Wetness Index (WI) is defined as follows (Burrough &

McDonnell 1998):

8/8/2019 15arspc Submission 19

9/12

9

)tan / ln( s AWI (eq.1)where A s is the upstream area for a cell and is the slope in degrees.

(a) (b)

(c)

Figure 8. a) GPS transects used for DEM creation; b) DEM interpolated from the transects; c) UAV photograph and location of moss quadrats draped over the DEM.

The wetness index is very sensitive for errors in the DEM as a small error inheight can significantly change the flow direction of a cell. To account for thesemeasurement and interpolation errors we applied a Monte Carlo simulation.This simulation assigns a random error to each grid cell in the DEM andcalculates the wetness index based on the new DEM (Revill et al. 2007;Burrough, van Gaans, & MacMillan 2000). The error is assumed to be normallydistributed with a mean of 0.0 and a user-defined standard deviation. Thisprocess is repeated 100 times resulting in 100 possible wetness index layers.

For each grid cell, the mean of these 100 realisations is taken to represent the

8/8/2019 15arspc Submission 19

10/12

10

final wetness index layer taking into account errors in the DEM. A combinedmeasurement and interpolation error of 0.2 m was used in this study, based onthe average GPS measurement error in height and an estimate for the stabilityof the handheld GPS antenna pole. A 3D model of the wetness index map is

shown in Figure 9 . Comparing Figure 8 c and Figure 9 indicates that there is avery strong relationship between water availability as modelled by the wetnessindex and the location and extent of the moss bed. In future studies we areplanning to concentrate on quantifying the relationship between moss healthand environmental conditions in more detail using UAV flights, fieldphotography, and additional spatial data collection.

Figure 9. 3D representation of the study sites in the Windmill Islands showing the

topographic wetness index derived from a digital elevation model (DEM). Dark areas have the highest water accumulation. Samples are positioned along hydrological gradients.

Conclusions

Antarctic moss beds are an important indicator for climate change. In recentyears a significant decline in bryophyte health has been observed. It is crucial tocollect detailed spatial and temporal information on these moss beds for reliablemonitoring. This study has demonstrated that a UAV is an effective tool forultra-high resolution mapping of moss beds in Antarctica. Aerial photographyacquired by the UAV has a spatial resolution of 1.5 cm, which is sufficientlydetailed to identify the location and extent of the moss beds. A detailed DEMwas generated based on GPS transects. A wetness index was extracted fromthe DEM to characterise the environmental conditions around the moss bed andto assess the effect of micro-topography. Future research will focus onquantification of the relationships between moss health and terrain information,based on a combination of UAV photography, field photography, and advancedDEM derivatives.

8/8/2019 15arspc Submission 19

11/12

11

Acknowledgements

We would like to thank the Australian Antarctic Division for research fundingand logistical support (project: AAS313). In addition, we would like to thank theCasey staff for field support. Finally, we thank Prof. Richard Coleman forfunding the thermal camera.

References

Berni, J.A., Zarco-tejada, P.J., Surez, L. & Fereres, E. (2009) Thermaland Narrowband Multispectral Remote Sensing for Vegetation Monitoring Froman Unmanned Aerial Vehicle. IEEE Transactions on Geoscience and Remote Sensing , 47 , 722-738.

Burrough, P.A. & McDonnell, R.A. (1998) Principles of Geographical Information Systems . Oxford University Press.

Burrough, P.A., van Gaans, P.F. & MacMillan, R.A. (2000) High-resolution Landform Classification Using Fuzzy k-means. Fuzzy Sets and Systems , 37-52.

Convey, P., Bindschadler, R., di Prisco, G., Fahrbach, E., Gutt, J.,Hodgson, D., Mayewski, P., Summerhayes, C. & Turner, J. (2009) Antarcticclimate change and the environment. Antarctic Science , 21 , 541.

Dunford, R., Michel, K., Gagnage, M., Piegay, H. & Tremelo, M.L. (2009)

Potential and constraints of Unmanned Aerial Vehicle technology for thecharacterization of Mediterranean riparian forest. International Journal of Remote Sensing , 30 , 4915-4935.

Hardin, P. & Jackson, M. (2005) An unmanned aerial vehicle forrangeland photography. Rangeland Ecology & Management , 58 , 439 442.

Hunt Jr, E., Hively, W., Fujikawa, S., Linden, D., Daughtry, C. & McCarty,G. (2010) Acquisition of NIR-Green-Blue Digital Photographs from UnmannedAircraft for Crop Monitoring. Remote Sensing , 2, 290 305.

Hutchinson, M.F. (1989) A new procedure for gridding elevation andstream line data with automatic removal of spurious pits. Journal of Hydrology ,106 , 211-232.

King, D. (2009) Rapid change in continental Antarctic vegetationcommunities.

Laliberte, A.S. & Rango, A. (2009) Texture and Scale in Object-BasedAnalysis of Subdecimeter Resolution Unmanned Aerial Vehicle (UAV) Imagery.IEEE Transactions on Geoscience and Remote Sensing , 47 , 761-770.

8/8/2019 15arspc Submission 19

12/12

12

Lelong, C.C., Burger, P., Jubelin, G., Roux, B., Labb, S. & Baret, F.(2008) Assessment of unmanned aerial vehicles imagery for quantitativemonitoring of wheat crop in small plots. Sensors , 8, 3557-3585.

Lovelock, C. & Robinson, S. (2002) Surface reflectance properties ofAntarctic moss and their relationship to plant species, pigment composition andphotosynthetic function. Plant, Cell & Environment , 25 , 1239 1250.

Nagai, M., Chen, T., Shibasaki, R., Kumagai, H. & Ahmed, A. (2009)UAV-Borne 3-D Mapping System by Multisensor Integration. IEEE Transactions on Geoscience and Remote Sensing , 47 , 701 708.

Rango, A., Laliberte, A., Steele, C., Herrick, J., Bestelmeyer, B.,Schmugge, T., Roanhorse, A. & Jenkins, V. (2006) Using Unmanned AerialVehicles for Rangelands: Current Applications and Future Potentials.Environmental Practice , 8, 159 168.

Revill, A.T., Snape, I., Lucieer, A. & Guille, D. (2007) Constraints ontransport and weathering of petroleum contamination at Casey Station,Antarctica. Cold Regions Science and Technology , 48 , 154-167.

Robinson, S.A., Wasley, J. & Tobin, A.K. (2003) Living on the edge plants and global change in continental and maritime Antarctica. Global Change Biology , 9, 1681 1717.

Turner, J., Colwell, S.R., Marshall, G.J., Lachlan-Cope, T.a., Carleton,A.M., Jones, P.D., Lagun, V., Reid, P.a. & Iagovkina, S. (2005) Antarctic climatechange during the last 50 years. International Journal of Climatology , 25 , 279-294.

Wasley, J., Robinson, S.A., Lovelock, C.E. & Popp, M. (2006) Some likeit wet biological characteristics underpinning tolerance of extreme waterstress events in Antarctic bryophytes. Functional Plant Biology , 33 , 443.

Zarco-Tejada, P.J. (2008) A new era in remote sensing of crops withunmanned robots. SPIE Newsroom , 2-4.

Zhou, G., Ambrosia, V., Gasiewski, A.J. & Bland, G. (2009) Foreword tothe Special Issue on Unmanned Airborne Vehicle ( UAV ) Sensing Systems forEarth Observations. IEEE Transactions on Geoscience and Remote Sensing ,47 , 687-689.