Project Information

Research Approach of ViWA

To achieve the project objectives, a remote sensing based monitoring approach is coupled with global agro-hydrological models in order to, for the first time, on a worldwide and daily basis determine the agricultural, environmental and societal water flows, water use efficiency and agricultural yields on a small scale (1 km²) (target 1). The approach used is shown schematically in Fig.2:

 

Fig.2: The scaling and simulation concept of target 1. Meteorological inputs dynamically downscaled from global reanalysis data drive a 225 member ensemble of hourly simulations representing the major crops and cultivation practices (fertilization, irrigation, cropping intensity) for each 1km pixel on the global cropland plus 1 member for each non-cropland. The simulated ensemble of growth curves is compared with high resolution (20m) COPERNICUS remote sensing data streams to determine actual water use efficiency, actual yield and virtual water content of crops for ~150 global test sites.
Fig.2: The scaling and simulation concept of target 1. Meteorological inputs dynamically downscaled from global reanalysis data drive a 225 member ensemble of hourly simulations representing the major crops and cultivation practices (fertilization, irrigation, cropping intensity) for each 1km pixel on the global cropland plus 1 member for each non-cropland. The simulated ensemble of growth curves is compared with high resolution (20m) COPERNICUS remote sensing data streams to determine actual water use efficiency, actual yield and virtual water content of crops for ~150 global test sites.

The results are used to drive coupled surface- and groundwater-models for selected large river basins to simulate blue and green water flows, to validate the results of the simulation using measured data and to quantify competition for water between agriculture, ecosystems and human water supply (see Fig.3).

Based on the data from this monitoring and simulation system the marginal productivities of water use in agriculture are determined using econometric methods (target 2). This for the first time allows investigating on a local scale the contribution of marginal water use to agricultural production. This information will be used to evaluate the economic trade-offs of change in the water allocation with a Computable General Equilibrium model (target 2). The same model compound is used to simulate global water flows, the international agricultural trade and the associated virtual water flows for different scenarios (targets 3 and 7).

 Fig.3: Considered water flows in the representative global test watersheds (see Fig.2) which are considered in the analysis of competing water demands
Fig.3: Considered water flows in the representative global test watersheds (see Fig.2) which are considered in the analysis of competing water demands

The model results are also used to evaluate the competition for water between industry, households, agriculture and natural or semi-natural ecosystems (target 6), to analyze the consequences of their interannual variability (e.g. El Nino, La Nina) (target 4) and to quantify the effects of different sustainability strategies to improve the local water situation and identify time-changing cold spots and hot spots of water scarcity through appropriate indicators (UN 2015, targets 5, 7). A continuous dialogue with stakeholders and project observers ensures that the tools are design and used with practical applicability in mind and scenarios are defined, simulated and analyzed in such a way that they identify practice-relevant solution options and control mechanisms for a more sustainable regional and global water use. The project will be carried out with actual data of 2015 to 2018 in order to study natural variability of the global real and virtual water flows related to agriculture. It specifically covers the extreme El Nino event of 2015/16 (with its severe consequences for Africa) and the likely la Nina of 2016/17. A total of 150 global Sentinel testsites will be used and the complete time series of available Sentinel 2 and selected Sentinel 1 data analyzed and assimilated into the agro-hydrological model to determine, on a local scale, water flows, agricultural yields and water use efficiency. The results from the testsites are scaled up to the global level. The massive computational load of image processing and model simulations are handled by SuperMUC, the High Performance Computing center of the Bavarian Academy of Science. The adaptation of the models for use in a supercomputing environment and their application on high-performance computers is also intended to provide information on where and how Big Data and Supercomputing can support both implementation and monitoring of Sustainable Development Goals.