Evaluation of MODIS products for the quantification of crop water stress in Piedmont

Drought periods are increasing their frequency as consequence of climate change and the agricultural sector is one of the most affected due to the water demand and growing request of agricultural commodities. Thus, water management for a sustainable agriculture is becoming challenging. In this context, the need to estimate soil evapotranspiration (ET) arises because it is critical for crop water stress assessment as well as hydrological modelling, irrigation scheduling, water management and weather forecasting. Since an accurate estimation of ET is difficult to reach across different spatial and temporal scales, a satellite-based product for this type of analysis proves to be very useful. The purpose of this study is to analyse the water stress in the Piedmont region in Italy, which in recent years has been increasingly affected by intense periods of drought, to which it is not usually subject. In addition, the study aims at comparing the performance of the satellite-based MODIS evapotranspiration operational product (M*D16A2) with a mathematical soil water balance model. This latter provides an estimation of crop evapotranspiration based on the Food and Agriculture Organization guidelines (FAO n.56). The time frame of the analysis includes the two-year period 2021-2022, as these years were perceived as impacted by below-average rainfall and higher temperatures compared to seasonal averages. The analyses are divided into two parts and are performed at two different scales. The first part is conducted at regional level, encompassing all agricultural areas within the Piedmont region. In this first part, the focus was on assessing crop water stress by evaluating ET during the summer months, which is the period when most seasonal crops are typically cultivated. The results reveal ET values lower than the expected averages, which is the reason why a subsequent analysis at a local scale was conducted. To perform this latter, uniformly cultivated maize areas were identified in the provinces of Turin and Alessandria. In these areas a comparison was conducted between the data collected and processed by MODIS and those calculated using the mathematical model. The estimation of ET through the mathematical model was performed using the maximum and minimum temperature data as well as precipitation data from ARPA. The results highlight a significant discrepancy between the MODIS data and the estimates from the mathematical model in some cases. In particular, an underestimation of the MODIS product was observed, reaching daily actual ET values of 1mm/day during growing period both at district and local scale. By analysing the potential ET (PET) values of the MODIS product, it appears to be in line with the modelled reference ET (ET0). However, the reduction of PET values to actual ET (calculated as the ET/PET ratio) is very high, with results of about 0.20. The underestimation of ET by MODIS data can be attributed to an indirect measurement of the parameters of aerodynamic resistance, canopy resistance, and soil evaporation, which are considered in the Penman–Monteith equation. In the MODIS algorithm, these parameters are indeed the result of intermediate algorithm calculations. This study aims to contribute to the evaluation of the quality and limiting aspects of the M*D16A2 satellite product on croplands, which has significant potential for large-scale ET estimation and being a ready-to-use product.