pyVPRM: A next-generation Vegetation Photosynthesis and Respiration Model for the post-MODIS era

Abstract. The Vegetation Photosynthesis and Respiration Model (VPRM) is a well-established tool to estimate carbon exchange fluxes between the atmosphere and the biosphere. The gross primary production (GPP) and respiration (R_eco) of the ecosystem are modelled separately at high spatial and temporal resolution using the satellite-derived Enhanced Vegetation Index (EVI) and Land Surface Water Index (LSWI), as well as meteorological variables for solar irradiance and surface temperature. The net ecosystem exchange (NEE) is calculated as the difference between the gross fluxes GPP and Respiration. VPRM is widely used as a biospheric flux model in atmospheric transport modeling, most often on scales ranging from city to continent, but also in studies of biospheric carbon budgets and their changes with climate extremes. Historically, satellite-based surface reflectances from the 500-m-resolution Moderate Resolution Imaging Spectroradiometer (MODIS) have been used to determine the EVI and LSWI. However, MODIS is reaching the end of its lifetime and will soon be decommissioned. Therefore, we present an updated version of VPRM, pyVPRM, which provides a software framework with a modular structure that can be used with various satellite products, land cover maps, meteorological data sources, and VPRM model parameterizations. Our tool naturally provides an interface to use satellite data from Sentinel-2, MODIS and VIIRS, as well as global high-resolution land cover classification maps from the Copernicus Dynamic Land Cover Collection 3 and ESA World Cover at 100 m and 10 m resolution, respectively. Neither product is static, hence dynamic changes of the land cover from year to year can be represented. Using Sentinel-2, ecosystem fluxes can be calculated at a resolution of up to 20 m, providing more accurate flux estimates in heterogeneous landscapes like croplands and allowing to resolve small-scale vegetation patches as common in urban areas. In contrast, VIIRS data are at the same resolution as MODIS, and thus provide for continuity once MODIS is discontinued, requiring only minor adjustments to the VPRM data preprocessing. In addition, pyVPRM improves the data handling, for example for snow-covered scenes. This paper presents the pyVPRM framework, discusses changes and improvements compared to previous VPRM implementations, and provides VPRM parameters for the European domain based on indices calculated from MODIS, Sentinel-2 and VIIRS using a new, wind-speed-optimized selection of eddy-covariance observations from 97 flux tower sites. Using pyVPRM and the new parameters we observe significant improvements in the estimation of the European carbon budget. The results are well conform with those from inversion studies.