Improving Climate Resilience of Agricultural Systems through the Development of Drought Vulnerability Curves

Drought is a significant threat for the agricultural sector since it is highly dependent on climate and water resource. As reported by FAO, about 83% of all drought-related losses between 2005 and 2015 were suffered by the agriculture. The estimation of drought impacts on crops is fundamental to foresee the effects that future droughts will have on the sector. In Italy, 50% of the country’s water withdrawal is allocated to agriculture according to ISTAT. Twenty percent of the arable land is irrigated on average at national level. In the Northern regions the picture is quite different since the percentage of irrigated land with respect to arable land is higher (53%, 42% and 37% of irrigate arable land in Lombardy, Veneto and Piedmont respectively). The three regions are part of the Po basin, which is the largest Italian agricultural area and accounts for 35% of the country’s agricultural production. The basin was struck by multiple droughts over the past years. Previous studies highlighted that the 2003-2008 events caused huge economic impacts on the agricultural sector, with the 2005-2007 period accounting for around 1.857B€ of losses. The development of quantitative models to establish a relationship between water deficit and crop yield losses can help in reducing losses by improving the water allocation mechanisms to provide water when it is actually needed. This study develops crop specific vulnerability curves tailored to the Po river basin context. The curves show the relationships between water deficit and yield losses during various crop growth stages (vegetative, flowering and yield formation). The Agricultural Production System sIMulator (APSIM) was used to reproduce the plant growth. The crop model has been implemented specifically to provide accurate predictions of crop production in relation to climate. The model was initialized with daily meteorological parameters (rainfall, average, maximum and minimum temperature, solar radiation) from the E-OBS dataset, a gridded weather dataset with a 10km spatial resolution available for Europe only. The dataset is based on the interpolation of observational data. The spatial coverage is good over the Po basin area. Soil texture was retrieved from the ISRIC dataset, while the agricultural practices (such as sowing date, fertilizer amount, etc) were derived from Lombardy region guidelines for the year 2020. At first reference yield for a specific season (the yield in the absence of any water stress during the entire growing season) was computed. Then, the reduced yield for the same season was derived introducing a water stress in a single growth stage by progressively reducing the precipitation amount during the growth stage. The yield reduction was expressed as one minus the ratio between the reduced yield and the reference yield. From the APSIM model crop water deficit for each season and each growth stage was derived. The relationship between yield reduction and water deficit was plotted to derive the vulnerability curves. Data points were fitted to asymmetric logistic functions. Two crops were considered: maize and winter wheat. The developed vulnerability curves for the selected crops are in agreement with the vulnerability functions proposed by the FAO, which has underlined the importance of avoiding water stress during flowering to get high yield. Vulnerability curves here developed represent a useful tool to support future decision-making strategies under climate risk conditions. Vulnerability curves can help improving resilience in agriculture, as they represent easy to use tools to warn farmers about climate vulnerability of their crops, and consequentially to support more resilient crop development and planning.