Estimation of actual evapotranspiration using remote sensing data
MetadataShow full item record
Gamage, Nilantha; Smakhtin, Vladimir; Perera, B. J. C. 2011. Estimation of actual evapotranspiration using remote sensing data. In Chan, F.; Marinova, D.; Anderssen, R. S. (Eds.). MODSIM 2011, 19th International Congress on Modelling and Simulation - Sustaining Our Future: Understanding and Living with Uncertainty, Perth, Australia, 12-16 December 2011. Canberra, Australia: Modelling and Simulation Society of Australia and New Zealand. pp.3356-3362.
Permanent link to cite or share this item: http://hdl.handle.net/10568/38462
Estimation of actual evapotranspiration (AET) and its spatial distribution are important to understanding of catchment hydrology. The AET is driven by net energy available to evaporate water from soil and vegetation surfaces, and to transpirate water from vegetation. However, estimating AET is difficult as the evapotranspiration process involves complex physical and biological processes. It is further complicated when there is lack of measured meteorological variables data which are required for estimation. These data are essential to quantify the availability of net energy and the aerodynamic effects of the evapotranspiration process. Remote sensing (RS) data, which are widely available and easily accessible than the measured ground data, can be used to estimate the availability of net energy for AET. However, still some measured ground data are required to quantify the aerodynamic effects on AET. In this study, remote sensing data and readily available climate datasets were used as inputs to an energy balance technique to estimate AET, as an alternative to the traditional ET estimation procedures, which require measured hydrometeorological data. The Macalister subcatchment in the Thomson catchment in Victoria (Australia) was used as the case study considering the study period from January 2003 to December 2008. Reflectance and radiance data of Moderate Resolution Imaging Spectroradiometer (MODIS) on Terra satellite were the primary source of RS data. Tilted and absent images of Terra MODIS were replaced with Aqua satellite data. First, vegetation indices such as Normalized Difference Vegetation Index, Leaf Area Index, fractional vegetation coverage and broadband albedo were calculated based on MODIS reflectance data for non-cloudy days. Similarly, MODIS radiance data were used to generate daily surface temperature on non-cloudy days. Vegetation indices and surface temperature were then used in the Surface Energy Balance System (SEBS) to estimate AET for non-cloudy days. However, SEBS requires data on a limited number of meteorological variables to quantify the aerodynamic effect of AET, and those data were obtained from ground measurements and global climate datasets (i.e. IWMI climate and water atlas). Once the AET was estimated for non-cloudy days, eostationary Operational Environmental Satellite (GOES) data were used to calculate fractional cloud cover and used to estimate net radiation available for cloudy day AET. The accuracy of the non-cloudy and cloudy day AET estimated using RS data was studied using root mean square error (RMSE) and Nash-Sutcliffe efficiency (Ef), by considering Penman-Monteith (PM) based AET as the observed AET, at four different sites in the atchment. A crop coefficient was used to convert PM based reference crop evapotranspiration to PM based AET. Remote sensing based AET shows higher coefficient of determination (R2) compared to PM based AET on non-cloudy days and comparatively less R2 on cloudy days. Results revealed that RS based AET overestimated during non-cloudy days, especially when the AET is more than 3 mm/day. However, RS based AET underestimated during partial and total cloud days. The above observations are common to all selected sites. Similar observations were seen with RMSE and Ef at all sites. The results show that remote sensing data and global climate dataset can be successfully used to estimate AET for the catchments where required ground measured meteorological data are not available. The estimated AET can be used as input to streamflow simulation models to generate streamflow in data poor catchments.
In Chan, F.; Marinova, D.; Anderssen, R. S. (Eds.). MODSIM 2011, 19th International Congress on Modelling and Simulation - Sustaining Our Future: Understanding and Living with Uncertainty, Perth, Australia, 12-16 December 2011. Canberra, Australia: Modelling and Simulation Society of Australia and New Zealand