Data acquired in WP1 will help to improve the modeling of how plants regulate their water use in response to drought stress. We will scale the STEMMUS-SCOPE model to the regional/national level by considering plant hydraulics and combine it with satellite observation of surface radiative and vegetation variables (albedo, surface temperature, leaf area, gross primary production) to simulate resilience of water-soil-plant systems in prolonged drought events. The processes and components of the STEMMUS-SCOPE system is illustrated in Figure 4. WP2 will also develop two emulated versions of STEMMUS-SCOPE with the physics-informed machine learning approach: a forward modeling emulator, utilizing original model input-output pairs, serving as a surrogate machine learning model to approximate the STEMMUS-SCOPE model; and an inverse modeling emulator, utilizing reflectance and fluorescence (together with original model input data) as predictor variables to train the emulator to predict above and belowground physiological variables.
Figure 4 Processes and components of the STEMMUS-SCOPE system