Although most part of the modeling in CGMS is based on WOFOST and LINGRA, other models are taken into consideration for tasks that fall outside the functionalities of CGMS. For this reason comparisons are made with the model from package, and models.
Integration of CGMS DBs and CERES-Wheat
A specific evaluation of the potential effects of frosts and cold temperatures on winter wheat was made integrating CGMS DBs and CERES-Wheat. The main purpose was to improve the frost-risk impact information especially on Central and Eastern European Countries where frost is a significant limiting factor for crop yield.
Since a dedicated subroutine to evaluate the effects of frost on winter crops is not currently available within the WOFOST version used in CGMS (Supit et al., 1994; Van Diepen, 2003) a module from CERES-Wheat (Ritchie, 1991b) was adapted to this analysis (Lazar et al., 2004). In order to integrate the CERES algorithms with the CGMS DBs ad-hoc routines in VBA have been developed. The daily outputs of this integration are the lethal temperature at the crown level based on the calculated hardening index (Gusta and Fowler, 1976????) and a coefficient for the reduction of the existing leaf area index.
In comparison with the previous maps for frost risks based on the occurrence of low temperatures in the areas with insufficient snow insulation, the new elaborations are additionally taking into account the dynamics of physiological status of the crop (hardening index) and provide separate insights on plant survival and damages of foliar apparatus. The data available from the CGMS Crop Knowledge Base, specifically from the MOCA project (Kucera and Genovese eds., 2004), were integrated in CERES-Wheat algorithms based routines resulting in a better understanding of the dynamics of the main factors explaining frost risks.
This tool allows a better estimation of the plant frost resistance taking into consideration its physiological condition (i.e. “hardening index”) which depends on the previous “thermal history” of the plant. The frosting conditions are determined as a combination of snow depth and low temperature. However, the plant-frost response (plant damages) to frost conditions may in fact vary according to the hardening index, which is determined by the pre-frost thermal conditions. This hardening index is used to calculate the killing temperature at crown level (about 3 cm below soil level, where the point for the initiation of new leaves is located). When the average temperature calculated for the crown level (depending mainly by snow layer depth) drops below the killing temperature level, severe plant population reductions are expected (high risk for crop failure). If only the crown minimal temperature is close to “killing” this must be considered as a risky situation (due to the field variations of the snow cover).Besides these immediate (lethal) effects, a daily cold dependent factor for the reduction of existing leaf area index (LAI) is calculated. Because the algorithm does not allow LAI reduction below a minimum level this calculated factor is used as an indicator for a possible delay in the spring “re-start” of vegetation. The model is dependent from a good simulation of the phenology – conditioned mainly by the use of the “correct” sowing date. The source of the crop calendar is twofold, the one proposed by Willekens et al. (1998)???? and the most recent proposed by Kucera (Kucera and Genovese eds., 2004) on the Central Countries. A user interface was designed to facilitate these operations. This approach can be optimized using cultivar specific parameters. The daily simulated values for development stage, hardening index, lethal temperature and reduction coefficient for the reduction of the LAI are stored in a GIS DB compatible with ARCVIEW and linked to the official boundaries (GISCO coverage). This system proved to be useful for detection of potential direct problems induced by low temperatures and also for quality check of snow data. An extensive testing with the historical weather data from 1975 to -2003 is on progress. Besides, the output from this application can be used as “predictors” into the Yield Forecasting.