Environmental constraints to net primary productivity at northern latitudes: A study across scales of radiation interception and biomass production of potato

Intercepted photosynthetically active radiation (Ipar, MJ m-2 d-1) is a key biophysical variable governing plant photosynthetic rate and net primary productivity (NPP, g m-2 d-1). Under optimal growth conditions, Ipar scales proportionally to NPP by a factor termed ‘optimum radiation use efficiency’ (RUEopt). The Carnegie-Ames-Stanford approach (CASA) considers temperature and moisture constraints to RUEopt and has been widely used in remote sensing studies to estimate productivity of various ecosystems. However, scale effects on the CASA have not been sufficiently investigated nor quantified. In this study, data at scales of field (Rapidscan reflectance data), air- (unmanned aerial vehicle (UAV) imagery) and spaceborne (Sentinel-2 imagery), as well as for several environmental constraints, were utilized to develop and test the CASA. The test plant was potato grown for multiple years on sandy soils in Denmark. The results showed that data from all scales provided comparable estimates of daily fraction of Ipar (fIpar) and Ipar. Maximum air temperature and diffuse radiation were the most important environmental factors constraining RUEopt of potato. Taking these into account considerably improved the prediction of NPP with the CASA (R2 increase of 25–32 % compared to no constraints), whereas stress effects due to soil moisture and vapor pressure deficits were less important (relative improvement in R2 of 2–3 %). Hence, RUEopt was 4.66, 4.19 and 4.98 g MJ-1 for Rapidscan, UAV and Sentinel-2, respectively, which was about 47 %–59 % higher than the estimates of the observed actual radiation use efficiency. This study offers an operational method for deriving RUEopt at plant species level since remote sensing and meteorological data are both readily available, and demonstrates that environmental constraints of the CASA does vary, depending on the study region. The method can be used to predict dry matter production in-season for other crops as well, with potential for extending the approach to determine fertilization and irrigation demands.