Operative Approach to Agricultural Drought Risk Management

In a companion paper, the agricultural drought economic risk assessment (ADERA) model was described. ADERA studies agricultural drought not only in terms of deficit soil water content and its frequency and severity (as in the traditional approach) but also in terms of net benefit reduction and vulnerability. The relationship between drought index, crop yield, and crop economic net benefit is modeled and three threshold levels are considered: the critical water content; the critical impact in terms of net benefit; and its critical return period (i.e., frequency) above which the soil-climate unit is considered unsuitable for the cultivation of a specific crop. These critical levels are used in long-term action planning as the triggers for different risk classes. The rapidity with which the critical impact is reached is considered indicative of the crop vulnerability, and real-time monitoring of the impact evolution (short-term actions) is also implemented. Initially ADERA was applied at only one point (soil-crop-climate unit), but because agricultural drought is by definition a wide area phenomenon, the model must be validated in other units and then used to evaluate the drought area extension (risk map). In this paper, generalization, demonstration, and guidance on the application of ADERA are given with reference to rainfed sunflower and sorghum crops in six soil-climate units in central Italy for which time series for yield (obtained in experimental plots where the water scarcity was the principal factor limiting the final yield) since 1978 is available. The application of ADERA made it possible to quantify the vulnerability class and the corresponding risk class for the single unit (long-term action planning). The analysis results confirm that typically for the selected crops the critical impact has a low risk and a low vulnerability. In fact sunflower and sorghum are rainfed crops widely cultivated in the area under investigation. Moreover, the results show that each risk class is present only within a specific cost range, in particular the upper and lower limits, and the size of each cost range varies according to the different units; the occurrence of a specific risk class and vulnerability level is highly correlated with the selected triggers, in particular with the critical net benefit (for an increment of which above the corresponding cost range, a given soil-climate unit can become unsuitable for the cultivation of a specific crop); the critical water content is a normal occurrence event for all the units analyzed, with the crop water stress onset date always around 40% of the growing period being a typical characteristic of the specific area, of the two crops and of the selected drought index. The practical usefulness of the impact monitoring procedure was stressed by comparing its progressive evolution in the cases of both a rainfed crop and of a rescue irrigation carried out at a selected and opportune date. In the latter case the impact was reduced below the critical level.