Improved consideration of PAR in vegetation carbon sink modelling using time series of remote-sensing data

When regarding the role of vegetation as a sink for atmospheric carbon, the manner in which applicable models consider the energy available for photosynthesis has been somewhat limited in its sophistication. Data collected from remote sensing satellites offer the possibility of more realistic estimations of Photosynthetically Active Radiation (PAR), which directly affects the level of increases in vegetative biomass. Geostationary satellites such as Meteosat Second Generation (MSG) will allow sub-hourly assessments of cloud cover and aerosol load. Thus it will be possible to obtain multiple estimates of PAR over the course of a day. The way in which Net Primary Production (NPP) models working on global or continental scales handle this important energy input varies. Some models require time integrated solar data as input (e.g., shortwave radiation, PAR), while others generate a daily cycle of PAR using changes in solar zenith angle while assuming constant atmospheric conditions (e.g., cloud cover, aerosol load). It is of great interest to try to answer questions regarding such sampling schemes in terms of accuracy, computing power and data availability as well as their impact on NPP calculations. In this study, a new technique to estimate PAR using satellite data combined with radiative transfer models is investigated. Initial results using AVHRR data to derive daily PAR estimates are presented, but the technique will be used to derive sub-daily estimates from MSG data as soon as available. To determine the best way to handle temporal sampling within an NPP model, investigations are performed with the BETHY model (Max-Planck-Institute of Meteorology, Hamburg) using different time resolutions of PAR input data.