Plant diseases affect the availability and safety of plants for human and animal consumption and threaten food safety, thus reducing food availability and access, as well as reducing crop yield and quality. There is a need for novel disease detection methods that can be used to reduce plant losses due to disease. Thus, this study aims to diagnose tomato leaf diseases by classifying healthy and unhealthy tomato leaf images using two pre-trained convolutional neural networks (CNNs): Inception V3 and Inception ResNet V2. The two models were trained using an open-source database (PlantVillage) and field-recorded images with a total of 5225 images. The models were investigated with dropout rates of 5%, 10%, 15%, 20%, 25%, 30%, 40%, and 50%. The most important results showed that the Inception V3 model with a 50% dropout rate and the Inception ResNet V2 model with a 15% dropout rate, as they gave the best performance with an accuracy of 99.22% and a loss of 0.03. The high-performance rate shows the possibility of utilizing CNNs models for diagnosing tomato diseases under field and laboratory conditions. It is also an approach that can be expanded to support an integrated system for diagnosing various plant diseases.

Drought forecasting plays a crucial role in drought mitigation actions. Thus, this research deals with linear stochastic models (autoregressive integrated moving average (ARIMA)) as a suitable tool to forecast drought. Several ARIMA models are developed for drought forecasting using the Standardized Precipitation Evapotranspiration Index (SPEI) in a hyper-arid climate. The results reveal that all developed ARIMA models demonstrate the potential ability to forecast drought over different time scales. In these models, the p, d, q, P, D and Q values are quite similar for the same SPEI time scale. This is in correspondence with autoregressive (AR) and moving average (MA) parameter estimate values, which are also similar. Therefore, the ARIMA model (1, 1, 0) (2, 0, 1) could be considered as a general model for the Al Qassim region. Meanwhile, the ARIMA model (1, 0, 3) (0, 0, 0) at 3-SPEI and the ARIMA model (1, 1, 1) (2, 0, 1) at 24-SPEI could be generalized for the Hail region. The ARIMA models at the 24-SPEI time scale is the best forecasting models with high R² (more than 0.9) and lower values of RMSE and MAE, while they are the least forecasting at the 3-SPEI time scale. Accordingly, this study recommends that ARIMA models can be very useful tools for drought forecasting that can help water resource managers and planners to take precautions considering the severity of drought in advance.

Climate, soil and tree water status, fruit yields and quality of ‘Valencia’ orange trees were monitored over five consecutive seasons (2007–2012) to study the effects of irrigation placement or volume. Three irrigation treatments were imposed: conventional irrigation (CI, 100% of crop evapotranspiration on both sides of the rootzone), partial rootzone drying (PRD, 50% of CI water only on one alternated side of the rootzone) and continuous deficit irrigation (DI, 50% of CI water on both sides of the rootzone). Yield parameters were generally not affected by PRD, and only yield per tree was lower in DI than CI trees. Fruit size and juice content were also reduced by DI, and not by PRD. Both PRD and DI increased juice soluble solids and acidity, vitamin C and carotenoid concentrations, as well as fruit, juice and sugar productivity per unit of irrigation water. Overall, the straight reduction of irrigation volumes by ~55% (DI) induced an average of 4.4% leaf dehydration, which caused significant fruit size reductions and a 3.4% reduction in juice yield, corresponding to a loss of about 2.6 tons of juice and 261 kg of sugars per hectare and year. On the other hand, water savings of about 2 mega liters per hectare and year with PRD caused a mild 2.3% leaf dehydration and mainly fruit quality improvements, indicating that PRD is a sustainable irrigation strategy for ‘Valencia’ orange.