This study evaluated the historical precipitation simulations of 49 global climate models (GCMs) of the Coupled Model Intercomparison Project Phase 6 (CMIP6) in reproducing annual and seasonal precipitation climatology, linear trends, and their spatial correlation with global SST across Africa and the Arabian Peninsula during the period of 1980–2014, using Global Precipitation Climatology Centre (GPCP) data as a reference. Taylor’s diagram was used to quantify the strengths and weaknesses of the models in simulating precipitation. The CMIP6 multi-mean ensemble (MME) and the majority of the GCMs replicated the dominant features of the spatial and temporal variations reasonably well. The CMIP6 MME outperformed the majority of the individual models. The spatial variation of the CMIP6 MME closely matched the observation. The results showed that at annual and seasonal scales, the GPCP and CMIP6 MME reproduced a coherent spatial pattern in terms of the magnitude of precipitation. The humid region received >300 mm and the arid region received <50 mm across Africa and the Arabian Peninsula. The models from the same modeling centers replicated the precipitation levels across different seasons and regions. The CMIP6 MME and the majority of the individual models overestimate (underestimate) in humid (arid and semi-arid)-climate zones. The annual and pre-monsoon seasons (i.e., DJFMA) were better replicated in the CMIP6 GCMs than in the monsoon-precipitation model (MJJASON). The CMIP6 MME (GPCP) showed stronger wetting (drying) trends in the northern hemisphere. In contrast, a strong drying trend in the CMIP6 MME and a weak wetting trend in the GPCP were shown in the Southern Hemisphere. The CMIP6 MME captures the spatial pattern of linear trends better than individual models across different climate zones and regions. The relationship between precipitation and sea-surface temperature (SST) exhibited a high spatial correlation (−0.80 and 0.80) with large variability across different regions and climate zones. The GPCP (CMIP6 MME) exhibited a heterogenous (homogeneous) spatial pattern, with higher correlation coefficients recorded in the CMIP6 MME than in the GPCP in all cases. Individual models from the same modeling centers showed spatial homogeneity in correlation values. The differences exhibited by the individual GCMs highlight the significance of each model’s unique dynamics and physics; however, model selection should be considered for specific applications.

Drought research is an important aspect of drought disaster mitigation and adaptation. For this purpose, we used the Standardized Precipitation Evapotranspiration Index (SPEI) to investigate the spatial-temporal pattern of drought and its impact on crop production. Using monthly precipitation (Precip) and temperature (Temp) data from 1986–2015 for 39 weather stations, the drought index was obtained for the time scale of 3, 6, and 12 months. The Mann–Kendall test was used to determine trends and rates of change. Precip and Temp anomalies were investigated using the regression analysis and compared with the drought index. The link between drought with large-scale atmospheric circulation anomalies using the Pearson correlation coefficient (R) was explored. Results showed a non-uniform spatial pattern of dryness and wetness which varied across Myanmar agro-ecological zones and under different time scales. Generally, results showed an increasing trend for the SPEI in the three-time scales, signifying a high tendency of decreased drought from 1986–2015. The fluctuations in dryness/wetness might linked to reduction crop production between 1986–1999 and 2005, 2008, 2010, 2013 cropping years. Results show relationship between main crops production and climate (teleconnection) factors. However, the low correlation values (i.e., <0.49) indicate the extent of the relationship within the natural variability. However, readers are urged to interpret this result cautiously as reductions in crop production may also be affected by other factors. We have demonstrated droughts evolution and trends using weather stations, thus providing useful information to aid policymakers in developing spatially relevant climate change adaptation and mitigation management plans for Myanmar.

Rwanda has experienced high temperature rising phenomena over the last decades and hence, highly vulnerable to climate change. This paper examined the spatial and temporal variations of daily maximum and minimum surface air temperature (Tmin and Tmax) and diurnal temperature range (DTR). It studied variables at monthly, seasonal and annual time-scales from 1961 to 2014. The study applied various statistical methods such as ordinary least-square fitting, Mann-Kendall, Sen’ slope and Sequential Mann-Kendall statistical test to the new reconstructed ENACTS dataset that cover the period from 1983 to 2014 while pre-1983s recorded data from 24 meteorological stations have been added to complete the lengthiness of ENACTS data. The January to February season did not show a significant trend at seasonal time-scales. The authors decided only to consider March-to-May, June-to-August and October-to-December seasons for further analyses. Topography impacts on temperature classified stations into three regions: region one (R1) (1,000–1,500 m), region two (R2) (1,500–2,000 m) and region three (R3) (≥2,000 m). With high confidence, the results indicate a significant positive trend in both Tmin and Tmax in all three regions during the whole study period. However, the magnitude rate of temperatures change is different in three regions and it varies in seasonal and annual scale. The spatial distributions of Tmax and Tmin represent a siginificant warming trend over the whole country notably since the early 1980s. Surprisingly, Tmin increased at a faster rate than Tmax in R3 (0.27 vs. 0.07°C/decade in March-to-May) and (0.29 vs. 0.04°C/decade in October-to-December), resulting in a significant decrease in the DTR. This is another confirmation of warming in Rwanda. The mutation test application exhibited most of the abrupt changes in the seasonal and annual Tmax and Tmin trends between 1984 and 1990. The present work mainly focus on the spatial and temporal variability of Tmin, Tmax and DTR in Rwanda and their relationship with elevation change, leaving a gap in other potential cause factors explored in the future.

Complex physical processes that are inherent to rainfall lead to the challenging task of its prediction. To contribute to the improvement of rainfall prediction, artificial neural network (ANN) models were developed using a multilayer perceptron (MLP) approach to predict monthly rainfall 2 months in advance for six geographically diverse weather stations across the Benin Republic. For this purpose, 12 lagged values of atmospheric data were used as predictors. The models were trained using data from 1959 to 2017 and tested for 4 years (2018–2021). The proposed method was compared to long short-term memory (LSTM) and climatology forecasts (CFs). The prediction performance was evaluated using five statistical measures: root mean square error, mean absolute error, mean absolute percentage error, coefficient of determination, and Nash–Sutcliffe efficiency (NSE) coefficient. Furthermore, Taylor diagrams, violin plots, box error, and Kruskal–Wallis test were used to assess the robustness of the model’s forecast. The results revealed that MLP gives better results than LSTM and CF. The NSE obtained with the MLP, LSTM, and CF models during the test period ranges from 0.373 to 0.885, 0.297 to 0.875, and 0.335 to 0.845, respectively, depending on the weather station. Rainfall predictability was more accurate, with 0.512 improvement in NSE using MLP at higher latitudes across the country, showing the effect of geographic regions on prediction model results. In summary, this research has revealed the potential of ANN techniques in predicting monthly rainfall 2 months ahead, supplying valuable insights for decision-makers in the Republic of Benin.