Meyer, H.; Lehnert, L.; Wang, Y.; Reudenbach, C.; Nauss, T. & Bendix, J. (2017): <b>From local spectral measurements to maps of vegetation cover and biomass on the Qinghai-Tibet-Plateau: Do we need hyperspectral information?</b>. <i>International Journal of Applied Earth Observation and Geoinformation</i> <b>55</b>, 21-31.
Resource Description
Title:
From local spectral measurements to maps of vegetation cover and biomass on the Qinghai-Tibet-Plateau: Do we need hyperspectral information?
FOR816dw ID:
281
Publication Date:
2017-03-15
License and Usage Rights:
PAK 823-825 data user agreement. (www.lcrs.de/dataagreementp3.do)
Resource Owner(s):
Individual:
Hanna Meyer
Contact:
email:
webmaster <at> lcrs.de
Individual:
Lukas Lehnert
Contact:
email:
lukas.lehnert <at> staff.uni-marburg.de
Germany
Individual:
Yun Wang
Contact:
email:
webmaster <at> lcrs.de
Individual:
Christoph Reudenbach
Contact:
email:
webmaster <at> lcrs.de
Individual:
Thomas Nauss
Contact:
email:
webmaster <at> lcrs.de
Individual:
Jörg Bendix
Contact:
email:
bendix <at> staff.uni-marburg.de
Deutschhausstraße 12
Room No. 02 A 48
35032 Marburg
Faculty of Geography
Germany
Abstract:
Though the relevance of pasture degradation on the Qinghai-Tibet Plateau (QTP) is widely postulated, its<br/>
extent is still unknown. Due to the enormous spatial extent, remote sensing provides the only possibility<br/>
to investigate pasture degradation via frequently used proxies such as vegetation cover and aboveground<br/>
biomass (AGB). However, unified remote sensing approaches are still lacking. This study tests the appli-<br/>
cability of hyper- and multispectral in situ measurements to map vegetation cover and AGB on regional<br/>
scales. Using machine learning techniques, it is tested whether the full hyperspectral information is<br/>
needed or if multispectral information is sufficient to accurately estimate pasture degradation prox-<br/>
ies. To regionalize pasture degradation proxies, the transferability of the locally derived ML-models to<br/>
high resolution multispectral satellite data is assessed. 1183 hyperspectral measurements and vegeta-<br/>
tion records were performed at 18 locations on the QTP. Random Forests models with recursive feature<br/>
selection were trained to estimate vegetation cover and AGB using narrow-band indices (NBI) as predic-<br/>
tors. Separate models were calculated using NBI from hyperspectral data as well as from the same data<br/>
resampled to WorldView-2, QuickBird and RapidEye channels. The hyperspectral results were compared<br/>
to the multispectral results. Finally, the models were applied to satellite data to map vegetation cover and<br/>
AGB on a regional scale. Vegetation cover was accurately predicted by Random Forest if hyperspectral<br/>
measurements were used (cross validated R2 = 0.89). In contrast, errors in AGB estimations were consid-<br/>
erably higher (cross validated R2 = 0.32). Only small differences in accuracy were observed between the<br/>
models based on hyperspectral compared to multispectral data. The application of the models to satellite<br/>
images generally resulted in an increase of the estimation error. Though this reflects the challenge of<br/>
applying in situ measurements to satellite data, the results still show a high potential to map pasture<br/>
degradation proxies on the QTP. Thus, this study presents robust methodology to remotely detect and<br/>
monitor pasture degradation at high spatial resolutions.<br/>
