This study aims to analyze the potential of different drought indices for identifying drought periods and predicting cereal yield in two semi-arid regions, Lleida in Catalonia and Kairouan in Tunisia, which have similar Mediterranean climates but different agricultural practices. Four drought indices, namely the Soil Moisture Anomaly Index (SMAI), the Vegetation Anomaly Index (VAI), the Evapotranspiration Anomaly Index (EAI), and the Inverse Temperature Anomaly Index (ITAI), were calculated from remote sensing data. Drought periods were identified from 2010/2011 to 2021/2022 based on the aforementioned indices. A correlation study between drought indices and wheat and barley yields was performed in order to select the most informative index and month for yield prediction. In the rainfed cereal area of Lleida, the strongest correlation was found between the EAI and VAI with barley yield (0.91 and 0.83, respectively) at the time of cereal maturity in June. For wheat, the strongest correlation was found between the EAI and VAI (0.75 and 0.72, respectively) at the time of cereal maturity in July. However, the VAI, EAI, and SMAI showed the best performance as an earlier indicator in March with a correlation with barley yield of 0.72, 0.67, and 0.64, respectively; the lowest standard deviation was for the SMAI. For wheat yield, the best earlier indicator was the SMAI in March, showing the highest correlation (0.6) and the lowest standard deviation. For the irrigated cereal zone of Kairouan, the strongest correlation (0.9) and the lowest standard deviation are found between the EAI and cereal yield in April. In terms of advanced prediction, the VAI shows a high correlation in March (0.79) while the SMAI shows a slightly lower correlation in February (0.67) and a lower standard deviation. The results highlight the importance of the EAI and SMAI as key indicators for the estimation and early estimation (respectively) of cereal yield.

This study developed a multi-year classification model for winter cereal in a semi-arid region, the Kairouan area (Tunisia). A random forest classification model was constructed using Sentinel 2 (S2) vegetation indices for a reference agricultural season, 2020/2021. This model was then applied using S2 and Landsat (7 and 8) data for previous seasons from 2011 to 2022 and validated using field observation data. The reference classification model achieved an overall accuracy (OA) of 89.3%. Using S2 data resulted in higher overall classification accuracy. Cereal classification exhibited excellent precision ranging from 85.8% to 95.1% when utilizing S2 data, while lower accuracy (41% to 91.8%) was obtained when using only Landsat data. A slight confusion between cereals and cereals growing with olive trees was observed. A second objective was to map cereals as early as possible in the agricultural season. An early cereal classification model demonstrated accurate results in February (four months before harvest), with a precision of 95.2% and an OA of 87.7%. When applied to the entire period, February cereal classification exhibited a precision ranging from 85.1% to 94.2% when utilizing S2 data, while lower accuracy (42.6% to 95.4%) was observed in general with Landsat data. This methodology could be adopted in other cereal regions with similar climates to produce very useful information for the planner, leading to a reduction in fieldwork.

With climate change, mainly drought, the situation of water stress in most Mediterranean countries is worsening with the high demand for agricultural water and the scarcity of water resources. Forecasts have found that more than 33 countries, including Tunisia and Spain, will face extremely high water stress by 2040, threatening agriculture and food security. In this study, we analyze the potential of different drought indices to identify drought periods for two regions with different climates: Kairouan in Tunisia, and Lleida in Spain, and we identify the indices that give more accuracy for cereal yield prediction.To achieve the objectives of this study, satellite data was used: MODIS (NDVI and LST) and SMOS. Spatial resolution enhancement algorithms have been applied, such as DISaggregation based on Physical And Theoretical scale Change (DISPATCh), to improve the spatial resolution of SMOS from 40 km to 1 km. In this study, we focus on two principal parameters to identify agricultural drought: Soil Moisture Anomaly Index (SMAI) calculated from soil moisture DISPATCh data, which gives an idea of the soil water status and Vegetation Anomaly Index (VAI) derived from MOD13Q1, which reflects the vegetative activity. Over the past 10 years, from the 2010/2011 agricultural year to 2019/2020, we have identified dry periods of agricultural drought based on VAI and SMAI. The results show that SMAI can detect more dry periods in space and time than VAI. For the study area in Tunisia, the strongest correlation obtained between wheat yield and SMAI is in November (R = 0.71). This result highlights the importance of water during this period. The correlation between wheat yield and SMAI decreased slightly in January (R=0.55), February (R=0.57), and March (R=0.63). However, the vegetation cover started to appear in January. A stronger, but later, correlation with VAI in March (R=0.63). For the second study area in Spain, Lleida, the correlation between drought index and yield anomaly of wheat and barley was studied separately. For barley, the increase in the correlation between grain yield and VAI started in February (R= 0.71), March (R=0.73), and then April where it reached its maximum (R=0.87). A more important correlation is noted in March with the SMAI which is about 0.8. Similarly, for wheat, the best correlation between yield and SMAI is recorded in March (R= 0.88) and with a slightly less important correlation with VAI of the order of 0.51 in March.In conclusion, this study shows the interest in improving the spatial resolution of soil moisture to better study agricultural drought and its effect on cereal yield.

Cereals represent an essential factor in the economy and food security, especially for countries that are not self-sufficient and depend on imports like Tunisia. Obtaining an early cereal map without the need to collect field data and without waiting for the end of the agriculture season helps the government to make early decisions. Hence, the first objective of our study is the development of an automatic classification model first calibrated for one agricultural year, 2020/2021 (2021), and then validated over the years 2011 through 2022 in the Kairouan governorate. The second objective is the development of a forecasting model in order to have early cereal maps several months before the harvest which occurs in June. Using Sentinel 2 and Landsat 5-7-8 data, different vegetation indices percentiles have been calculated. In order to select the best indices for cereal classification, a feature importance study over all the indexes was performed using the random forest classification algorithm reference year classification. A land cover classification model was validated for the reference year 2021, with an overall accuracy of 89.3%. This classifier has been used to elaborate land cover classification maps since 2011, focusing mainly on cereal crops. Using Sentinel 2 data, a good precision (P) for cereal crops was found, between 85,8% and 95,1%. Good to moderate accuracies were obtained when using Landsat data, between 41% and 91,8%. Then, a land cover forecasting model was validated for 11 years for different forecasting periods where we found excellent results four months before harvest (in February). We were able to obtain the cereal crop maps with a P between 85,1% and 95,1% using Sentinel 2 data and between 42,6% and 95,4% using Landsat data from four months before harvest. However, confusion between cereals and cereals grown with arboriculture was found which is due to the similarity between these two classes.With this automatic land cover model, we have been able to produce the cereal maps of the last 12 agricultural years. This approach could be also used in the future to obtain a cereal map as early as February.