Distributional Shifts Underestimate Net Primary Productivity in grazing grassland
▶Summary
Accurately estimating aboveground net primary productivity (ANPP), the key component of net primary productivity (NPP) and the primary food source for herbivores, is essential for improving our understanding of the carbon cycle. It is also critical for shaping future policy decisions and advancing sustainable grazing practices, especially in the context of climate change. However, the widespread use of inaccurate ground-truth data in the training and validation of ANPP models has led to problems with data distributional shifts, which reduce the accuracy of these models. In this project, “Distributional shifts Underestimate Net Primary Productivity in grazing grassland” (UeNPP), I will investigate the underlying mechanisms driving the data distributional shift in the simulation of grassland ANPP and provide solutions to improve grassland ANPP modelling accuracy. To achieve this, I will 1) investigate the components of ANPP in different grazing intensities across varied grassland types and evaluate the accuracy of the widely used methods for ANPP model training and validating. 2) identify changes in data distributional shifts over training, testing, and deployment phases. 3) provide two approaches, distributional shifts mitigation and standing-consumed biomass combination, to improve grassland ANPP modeling accuracy. and 4) apply the developed models to quantify the ANPP underestimation at the global scale. The MSCA-European Postdoctoral Fellowship will be hosted by the University of Copenhagen (UCPH, Denmark), home to world-leading experts in spatial monitoring and analysis, and supervised by Prof. Rasmus Fensholt, a renowned expert in Remote sensing/Earth observation ecology. The results of UeNPP will contribute to a better understanding of grassland biomass dynamics and improve decision-making processes for grassland managers globally. The findings will support the development of sustainable grazing practices and contribute to climate change mitigation efforts.