Improved global scale modelling of potential and actual evapotranspiration
Evapotranspiration plays a major role in the water cycle for catchment hydrology as well for global water circulation. In many areas most of the water is evapotranspirated before it reaches the river. A plausible estimation of (actual) evapotranspiration is therefore a prerequisite for hydrologic modeling (Beven 2001; Viessman Jr. & Lewis 2003). On the other hand, evapotranspiration is hard to measure.
The WaterGAP Global Hydrological Model (WGHM) computes the water balance of around 60000 land grid cells (each 0.5° x 0.5°) and routes the water through a drainage direction map until it reaches the outlet (mainly oceans). The model is calibrated on long term averages of discharge at 1235 stations. Various temporal and spatial data sets are required to run the model. Out of previous work, it seems that in large regions the water loss through the evaporation path is too low; resulting in too much modeled discharge in comparison to measured discharge. Therefore the calculation of evapotranspiration can be seen as a critical point in the current model approach. The three main steps of estimating the water loss through evaporation are i) calculation of net radiation, ii) calculation of potential evapotranspiration (PET) and iii) deriving the actual evapotranspiration (AET) out of PET and available (soil) water. Addressing these modeling steps, the following points will be the main aspects of this dissertation project.
- Climate input data (e.g. temperature, precipitation, cloudiness) are substantial drivers for calculating the radiation and water balance. Recently published climate data (CRU TS 3.1; H08) will be evaluated regarding their effect on the model results (see reference information). Especial the H08 dataset might have benefits in day-to-day variations of the water balance as it is a daily dataset.
- Spatial related information (esp. land cover, soil type) and its attributes (e.g. albedo, emissivity, rooting depth, water storage capacity) are influencing the calculation of radiation, PET and AET in different ways. The attributes currently used will be critically reviewed. The effect of variations in individual attributes will be investigated using a sensitivity analysis approach (mainly for radiation and PET). Additionally, new spatial information (e.g. soil map) and attributes will be reviewed in terms of enhancing the quality of currently used input.
- The estimation of potential evapotranspiration (PET) can be done using various approaches but is hard to verify. More physically based equations for estimating PET requires data which are hardly available in a good quality and on that scale whereas simple equations (e.g. Priestley-Taylor) are strongly dependent on some key parameter (as α). Main aspect will be to compare of the model outputs with currently published data (see reference information) and the optimization of parameters to get a realistic PET.
- AET is directly related to PET and water availability (e.g. soil water content). It will be investigated, how the currently used method can be optimized and if there are other approaches usable to represent this relationship.
The aspects described above are neither complete nor strict sequentially but can be seen as a field of work to improve the modeling of global potential and actual evapotranspiration.
A main disadvantage in previous attempts to improve PET and AET were the lack of global and temporally resolved data for comparison. Since the last years, many efforts were made by creating global dataset, either based on remote sensing techniques, combinations of measured data and regionalization approaches, and modeling activities. For example, Jimenez et al. (2010) compares 12 land surface heat flux estimates. The WATCH Model Intercomparison Project provides a range of different global scaled model results. Additional, (global) scientific networks were established in this area (e.g. FLUXNET). Even though there is no absolute truth out of these products and model outputs (especial PET), they can provide valuable reference information for improving the calculations of WGHM.
Primarily, the model success will be marked on the ability to simulate the river discharge within the range of measured discharge. Depending on the issue currently investigated, the spatial focus will vary from grid cell to river basins and global scale. Here, further information (e.g. literature based water balances of catchments, point estimates of radiation, PET and AET) will be taken into account as reference values. A high potential for improvement of the model is the inclusion of experiences from other environmental models. Based on single catchments, the behavior of storages and amounts of flows can be investigated comparatively to learn from each other.
Keywords:evaporation, transpiration, radiation, global hydrology
Participants: Hannes Müller Schmied, Milena Dopychai, Petra Döll
Duration: until Jan 2014
Funding entity:Hessian State Ministry for Science and Arts
Contact: hannes.mueller.schmied@em.uni-frankfurt.de
