PHASE: A geostatistical model for the Kriging-based spatial prediction of crop phenology using public phenological and climatological observations

摘要: Abstract Detailed information on plant developmental stages, referred as phenological phases, can assist research, applications and synergies e.g., in land use, climate science remote sensing. Usually, detailed ground about phases is only available point observations. However, most application scenarios of spatially interpolated required. In this article, we present an approach for modeling interpolation crop temperate climates the example total area Germany using statistical analysis a Kriging prediction process. The presented model consists two major parts. First, daily temperature observations are to retrieve countrywide data set. Second, linked day year which event was observed by governmental observation network. accumulated sum between sowing events calculated. any location exceeds phase-specific critical sum, indicates entry modeled phase, finally set specific phase. applied eight agricultural species including cereals, maize root crops 37 corresponding 2011. results tested show significantly lower mean squared errors (RMSE) values higher goodness fit ( R 2 ) compared computed Ordinary (OK) Inverse Distance Weighting (IDW). accuracy varies 2.14 days 11.45 days heading emergence winter wheat, respectively. uncertainty majority less than week. universally applicable due automatic parametrization, but accuracies depend type increase during growing season. possibility enhance additional explaining variables demonstrated consideration soil moisture within extended setting.