Search

Projecting future climate variables is essential for comprehending the potential impacts on hydroclimatic hazards like floods and droughts. Evaluating these impacts is challenging due to the coarse spatial resolution of global climate models (GCMs); therefore, bias correction is widely used. Here, we applied two statistical methods—standard empirical quantile mapping (EQM) and a hybrid approach, EQM with linear correction (EQM‐LIN)—to bias correct precipitation and air temperature simulated by nine GCMs. We used historical observations from 20 weather stations across South Florida to project future climate under three shared socioeconomic pathways (SSPs). Compared to the EQM, the hybrid EQM‐LIN method improved R2 of daily quantiles by up to 30% over the historical period and improved MAE up to 70% in months that contain most extreme values. Projected extreme precipitation at the weather stations showed that, compared to the EQM‐LIN, the EQM method underestimates the high quantiles by up to 26% in SSP585. The projected changes in annual maximum precipitation from historical period (1985–2014) to near future (2040–2069) and far future (2070–2100) were between 2% and 16% across the study area. Projected future precipitation suggested a slight decrease during summer but an increase in fall. This, along with rising summer temperatures, suggested that South Florida can experience rapid oscillations from warmer summers and increased flooding in fall under future climate. Additionally, our comparative analyses with globally and nationally downscaled studies showed that such coarse scale studies do not represent the climatic extremes well, particularly for high quantile precipitation. Plain Language Summary: Natural hazards, such as flooding and extreme heat, can damage our infrastructure and local communities. These hazards often stem from extreme weather conditions such as heavy precipitation and hot temperatures. Global climate change is expected to affect precipitation and air temperature, leading to more extreme weather events and hazards in the future. Predicting frequency and severity of these future events is key to protect our infrastructure and communities. Climate models have been widely used to project statistics of future extremes. However, they come with uncertainties due to numerical approximations, low spatial resolutions etc. Consequently, predictions generated by these models are inherently uncertain. This paper used a new approach to improve future projections of extreme climatic events, focusing on South Florida. Our findings revealed substantial improvements in projecting future events, particularly for extreme precipitation. We also showed that nationally and globally derived simulations may not be suitable for accurately projecting heavy precipitation, but can suffice for air temperature and low/medium precipitation rates. This paper offers promising avenues for refining projections of future weather events. By enhancing our ability to anticipate upcoming weather events more accurately, we can better protect our infrastructure and communities against the challenges posed by climate change. Key Points: A hybrid statistical approach to improve bias correction of GCM precipitation and air temperature simulationsMajor improvements in both climatic variables, especially extreme precipitation, using our hybrid statistical approach in South FloridaCoarse resolution downscaled data may be insufficient for deriving high quantile precipitation at the regional scale [ABSTRACT FROM AUTHOR]

[Objective] Understanding and grasping the temporal and spatial changes in meteorology and reasonably predicting the future trend of climate change can provide an important theoretical basis for the formulation of regional response measures. [Methods] The Sichuan section of the Yangtze River Basin was selected as the research area. Data analysis was conducted based on measured station data from 1980 to 2014. Three future hypothetical climate scenario models (SSP1-2.6, SSP2-4.5, SSP5-8.5) were estimated using the GCM with complete scenarios in the sixth phase of the International Coupled Model Comparison Plan (CMIP6) after bias correction, and the changes in precipitation climate variables within the region were analyzed. [Results (1) There is a certain degree of underestimation of precipitation simulation in climate models, with monthly scale seasonal scale annual scale in terms of time scale. (2) After the calibration and evaluation of the climate model, the data are more consistent with the actual situation, and the quantile mapping method is very effective for the calibration of the Climate model data. (3) Compared to the benchmark period (1980-2014), the future precipitation in the study area is higher than the historical period under three scenarios, with SSP5-8.5 SSP2-4.5 SSP1-2.6. (4) The areas with significant increases in precipitation are mainly located in the northwest of the study area, showing a spatial pattern of low level in the west and high level in the east. [Conclusion] The climate model's prediction of future precipitation changes in the study area shows an overall increasing trend, but there is underestimation to a certain extent. [ABSTRACT FROM AUTHOR]

Recognizing the differential impacts of climate change across geographical scales, this study emphasizes the importance of statistical downscaling. Using Gene Expression Programming (GEP) and Linear Genetic Programming (LGP), statistical downscaling transforms broad climate trends into region-specific insights. This allowed for detailed analyses of anticipated changes in sediment yield and discharge within a Euphrates River sub-basin in Türkiye using large-scale variables from the CanESM2 model. The dataset is divided into calibration (1970–1995) and validation (1996–2005) periods. To assess the models’ accuracy, statistical measures such as RMSE, MAE, NSE, and R were used. The analysis revealed that LGP outperformed GEP in both discharge and sediment yield during validation, with RMSE = 51.79 m3/s and 4,325.66 tons/day, MAE = 27.14 m3/s and 1,593.34 tons/day, NSE = 0.684 and 0.627, and R = 0.841 and 0.788, respectively. However, when simulating future periods based on the observed period (2006–2020), the GEP model was superior to LGP under RCP2.6, RCP4.5, and RCP 8.5 scenarios from CanESM2. In 2021–2100, models suggest a moderate decrease in discharge and sediment yield, indicating potential shifts in the basin's hydrodynamics. These changes could disrupt hydropower generation, challenge water management practices, and alter riverine ecosystems. The results necessitate a thorough assessment of potential ecological consequences. HIGHLIGHTS Climate change effects on streamflow and sediment yield have been explored.; Genetic programming such as GEP and LGP have been employed in statistical downscaling technique.; Future projection of streamflow and sediment yield have been done under different emission scenarios.;

This study examines two downscaling techniques, convolutional neural networks (CNNs) and feedforward neural networks for predicting precipitation and temperature, alongside statistical downscaling as a benchmark model. The daily climate predictors were extracted from the European Center for Medium-range Weather Forecast (ECMWF) ERA5 dataset spanning from 1979 to 2010 for Tabriz city, located in the northwest of Iran. The biases in precipitation data of ERA5 predictors were corrected through the empirical quantile mapping method. Also, two nonlinear predictor screening methods, random forest and mutual information, were employed, alongside linear correlation coefficient. While these methods facilitate identification of dominant regional climate change drivers, it is essential to consider their limitations, such as sensitivity to parameter settings, assumptions about data relationships, potential biases in handling redundancy and correlation, challenges in generalizability across datasets, and computational complexity. Evaluation results indicated that CNN, when applied without predictor screening, achieves coefficient of determination of 0.98 for temperature and 0.71 for precipitation. Ultimately, future projections were employed under two shared socioeconomic pathways (SSPs), SSP2-4.5 and SSP5-8.5, and concluded that the most increase in temperature by 2.9 °C and decrease in precipitation by 3.5 mm may occur under SSP5-8.5. HIGHLIGHTS Convolutional neural networks (CNNs) and feedforward neural networks (FFNNs) were used for downscaling general circulation model.; The empirical quantile mapping method was used for bias correction.; Future projections were employed under two shared socioeconomic pathways (SSPs), i.e., SSP2-4.5 and SSP5-8.5.; The results show the superiority of CNN over other AI methods.;

The impacts of climate change increasingly show themselves in many forms in our everyday lives such as heatwaves and droughts. Drought is one of the critical events today for increasing drought frequency. This study focuses on meteorological drought because it directly affects other drought types. Hence, this study focuses on how the future drought conditions will vary under climate change effects in the Mediterranean basin (Turkey). In doing so, this study utilizes precipitation data from three General Circulation Models (GCMs) and three Regional Circulation Models (RCMs). The GCMs are CNRM-CM6, GFDL-CM4, and MPI-ESM1, while the RCMs are (RCA4)-CNRM-CM5, (Reg CM4)-GFDL-ESM2M, and (RCA4)-MPI-ESM-MR. Mitigating biases of the climate models, this study utilizes four statistical downscaling methods (SD), linear scaling (LS), local intensity scaling (LOCI), power transformation (PT), and distribution mapping (DM). Here, the study has two purposes. The main aim of the paper here is to compare the performance of SD methods in improving the representation of observed climate variables in climate models. In addition, the study shows how different methods will affect the spatial drought distribution in the area under the SSP2 4.5 and SSP5 8.5 scenarios. Consequently, the study uses the standardized precipitation index (SPI) and Z-score index (ZSI) to quantify future drought conditions and reaches the following results. The study reveals that mild drought conditions are prevalent in the basin for future periods, and drought indices go down to − 0.55. The study also shows that different SD methods affect the results obtained by each climate model diversely. For example, while the LS method causes the most drought conditions on the results based on CNRM-CM5 and CNRM-CM6, the DM method has a similar impact on outcomes based on GFDL-CM4 and GFDL-ESM2M and causes the most drought conditions. [ABSTRACT FROM AUTHOR]

The Three‐River Headwater Region, also known as China's water tower, is highly sensitive to climate change and has experienced profound hydrological alterations in the last few decades. This study assessed the potential impacts of climate change on all the important hydrological components such as precipitation, evapotranspiration, streamflow, snow‐melt flow, and soil moisture (SM) content in the region. For this, climate data (i.e., temperature, precipitation, relative humidity, and windspeed) of three Global Climate Models (i.e., CanESM5, MPI‐ESM1.2‐HR, and NorESM2‐MM) was downscaled with the Statistical DownScaling Model (SDSM) and their ensemble was forced into a hydrological model to simulate the hydrological processes for 1981–2100. The screening process, which is central to all downscaling techniques, is very subjective in the SDSM. Therefore, we developed a quantitative screening approach by modifying the method applied by Mahmood and Babel (2013, https://doi.org/10.1007/s00704‐012‐0765‐0) for the selection of a set of logical predictors to cope with multi‐collinearity and their ranking. The analyses were performed for the near future period (NFP, 2021–2060) and far future period (FFP, 2061–2100) relative to the baseline period (BLP, 1981–2020). The results showed that the region will be hotter and wetter in the future, with intensive and frequent floods. For example, temperature, precipitation, evapotranspiration, and streamflow will increase by 1.0–1.5 (1–1.9)°C, 9–21 (15–27)%, 6–17 (9–29)%, and 9–46 (22–64)% in the NFP and by 2.0–2.8 (2.7–4.6)°C, 16–40 (43–87)%, 11–31 (24–73)%, and 20–95 (60–198)% in the FFP, respectively, under SSP2‐4.5 (SSP5‐8.5). Similar projections were explored for other hydrological components. Among all, surface flow showed an unprecedented increase (500%–1,000%) in the FFP. Peak flows will be much higher and will shift forward, and snowmelt will start earlier in the future. The results of the present study can be a good source for understanding the hydrological cycle and be used for the planning and management of water resources of the highly elevated and complex region of the Qinghai Tibetan Plateau. Plain Language Summary: The Three‐River Headwater Region, which is also known as the Sanjiangyuan in Chinese, is located in Qinghai Tibetan Plateau, China. It is considered the water tower of China because it is the source of three giant rivers the Yangtze, Yellow, and Lancang (Mekong). However, its water resources (hydrological cycle) are very sensitive and vulnerable to changing climate. Therefore, we assessed the potential impact of climate change on all the important hydrological components such as precipitation, streamflow, snow melt flow, surface flow, baseflow, soil moisture (SM), and changes in terrestrial water storage. Previous studies mainly focused on precipitation, streamflow, and SM. Global Climate Models (GCMs) are the main tool to assess the future changes in hydrological components under changing climate. Since GCMs have a coarse spatial resolution and biases in their outputs, a statistical downscaling model (SDSM) was applied to fix these issues and used to generate climate data (e.g., temperature and precipitation) for the future (2021–2100) under two scenarios (i.e., SSP2‐4.5 and SSP5‐8.5). These scenarios represent the global development and greenhouse gas emissions in the future. SSP2‐4.5 scenario typically involves moderate greenhouse gas emissions reduction efforts and some adaptation and mitigation measures to address climate change impacts, and SSP5‐8.5 represents high greenhouse gas emissions and limited efforts to mitigate climate change impacts. The screening process, which is central to all downscaling techniques, is very subjective in the SDSM. Therefore, we developed a quantitative screening approach by modifying the method applied by Mahmood and Babel (2013, https://doi.org/10.1007/s00704‐012‐0765‐0) for the selection of a set of logical predictors to cope with multi‐collinearity and their ranking. The downscaled future climate data was used as input to run a hydrological model (HEC‐HMS) to generate hydrological components under both scenarios. The future changes in the hydrological components were obtained for 2021–2060 and 2061–2100 with respect to the baseline period 1981–2020. The results showed that the region will be hotter and wetter in the future, with intensive and frequent floods. Almost all components are expected to increase in the future under both scenarios. Among all, surface flow showed an unprecedented increase (500%–1,000%) in the second half of the twenty‐first century (2061–2100). Peak flows are expected to be much higher than the present conditions and to shift forward. Snowmelt will start earlier in the future. This study will be very useful in understanding the hydrological cycle and can be used by policymakers, planners, and stakeholders for proactive adaptation strategies such as water resources planning and management, investments in water infrastructure, land use planning, ecosystem restoration, and community resilience‐building initiatives to mitigate potential risks. Key Points: The region will be hotter and wetter, with intensive and frequent floodsThe hydrological components are expected to increase in the futureSurface flow showed an unprecedented increase of 500%–1,000% [ABSTRACT FROM AUTHOR]

We analyzed hydrological responses to changing climate and land use/land cover (LULC) for the past (1985–2020) and future (2021–2080) in the Chemoga watershed of the Upper Blue Nile Basin. The watershed comprises four agroecological environments: Moist Kolla, Moist Weyna Dega, Moist Dega, and Wet Wurch. Past and projected LULC changes under business-as-usual (BAU) and land conservation (LC) scenarios were utilized. Climate projections from 2021 to 2080, under two Shared Socioeconomic Pathways (SSP2-4.5 and SSP5-8.5), were downscaled from Global Climate Models. Utilizing the Soil and Water Analysis Tool, we assessed impacts on mean annual surface runoff (SR) and evapotranspiration (ET). Maximum and minimum temperatures increased significantly in the past and future climate scenarios, with a significant rainfall increase observed under SSP5-8.5. Historical trends revealed a 16.6% increase in SR and 7% in ET from 1983–2002 to 2003–2020. Under BAU LULC with the SSP2-4.5 (SSP5-8.5) climate scenario, SR increased by 24% (26.1%) and ET by 3.1% (4.4%) from 2003–2020 to 2021–2050, followed by a subsequent SR rise of 13.7% (14.0%) and ET increase of 6.0% (5.7%) from 2021–2050 to 2051–2080. Conversely, the LC LULC with SSP2-4.5 (SSP5-8.5) resulted in a 5.3% (4.2%) SR decrease and ET increase of 9.7% (11.3%) from 2003–2020 to 2021–2050 and a further SR decrease of 1% (0.7%) and 6.1% (6.9%) ET increase from 2021–2050 to 2051–2080. The Moist Kolla agroecology experienced the highest SR increase due to vegetation clearances for commercial farming. Meanwhile, the LC scenario indicated substantial decreases in SR and marginal increases in ET in the Moist Weyna Dega agroecology due to forest restoration on steep slopes. Overall, SR showed greater sensitivity to LULC changes, while ET was more responsive to climate changes. The results emphasize considering diverse agroecological contexts for effective water resource management under changing climate and LULC scenarios. [ABSTRACT FROM AUTHOR]

Rainfall is a climate element with high variations in space and time scales, so it is not easy to predict. One way to predict rainfall is statistical downscaling (SD). SD can predict local rainfall based on Global Circulation Model (GCM) data. The Decadal Climate Prediction Project (DCPP), one of the GCMs, originates from adjacent grids and experiences multicollinearity problems. Rainfall as a response variable is Tweedie Compound Poisson Gamma (TCPG) distribution data because it has a discrete component (rainfall events) and a continuous component (rainfall intensity), so SD modelling will be carried out using Tweedie-LASSO. This research aims to compare the performance of bias correction and ensemble methods in SD in predicting rainfall in West Java, Indonesia. Bias correction uses Empirical Quantile Mapping (EQM) with CHIRPS data, and the ensemble method uses a stacking technique with Random Forest (Stacking-RF) due to the varied characteristics of DCPP model sources. Evaluation results using Root Mean Square Error Prediction (RMSEP) and correlation coefficient show that bias correction improves single-model performance but not ensemble models. Besides that, ensemble models outperform single models both before and after bias correction. The combination of bias correction and ensemble modelling can be recommended when conducting SD to enhance the prediction capability of rainfall at stations and other areas.

Abstract Projecting future climate variables is essential for comprehending the potential impacts on hydroclimatic hazards like floods and droughts. Evaluating these impacts is challenging due to the coarse spatial resolution of global climate models (GCMs); therefore, bias correction is widely used. Here, we applied two statistical methods—standard empirical quantile mapping (EQM) and a hybrid approach, EQM with linear correction (EQM‐LIN)—to bias correct precipitation and air temperature simulated by nine GCMs. We used historical observations from 20 weather stations across South Florida to project future climate under three shared socioeconomic pathways (SSPs). Compared to the EQM, the hybrid EQM‐LIN method improved R2 of daily quantiles by up to 30% over the historical period and improved MAE up to 70% in months that contain most extreme values. Projected extreme precipitation at the weather stations showed that, compared to the EQM‐LIN, the EQM method underestimates the high quantiles by up to 26% in SSP585. The projected changes in annual maximum precipitation from historical period (1985–2014) to near future (2040–2069) and far future (2070–2100) were between 2% and 16% across the study area. Projected future precipitation suggested a slight decrease during summer but an increase in fall. This, along with rising summer temperatures, suggested that South Florida can experience rapid oscillations from warmer summers and increased flooding in fall under future climate. Additionally, our comparative analyses with globally and nationally downscaled studies showed that such coarse scale studies do not represent the climatic extremes well, particularly for high quantile precipitation.

High-resolution bioclimatic data are crucial to providing fine-scaled insights into biodiversity assessment, forestry, and agricultural management. Existing global bioclimatic datasets often exhibit kilometer-level coarse resolution or have miss the data in recent decades, potentially resulting in the issues of lower spatial accuracy, limited information, and restricted applicability in fine-scaled studies. Hubei Province in Yangtze River Middle Reaches has sparse meteorological stations in high-altitude mountainous areas to map the high-resolution bioclimatic variables directly. This study developed a 30-year averaged bioclimatic dataset for Hubei Province during 1991–2020 at a 30-m spatial resolution, utilizing monthly temperatures and precipitation derived from a downscaling-calibration framework. The downscaling of 1-km resolution climate variables was achieved by using a random forest model with 30-m resolution terrain and spatial covariates. Then, the geographical differential analysis was applied to improve the accuracy of downscaled products by including additional ground data. The mean absolute errors of calibrated monthly maximum, mean, minimum temperatures, and precipitation based on ordinary kriging decreased from 0.74 °C, 0.47 °C, 0.47 °C, and 28.27 mm to 0.43 °C, 0.28 °C, 0.36 °C, and 21.43 mm, respectively. Finally, calibrated climate variables were employed to calculate 19 annual bioclimatic variables, which were subsequently averaged over the 30-year period. The constructed bioclimatic dataset exhibits high overall consistency with the WorldClim dataset according to pixel-based comparison (Spearman correlation coefficients >0.6), with differences mainly attributed to the superior local accuracy of our dataset and climate changes. The dataset will provide fine-scaled, updated, and reliable data supports for local-related studies and decision making.

Convolutional neural networks (CNNs) have become one of the state‐of‐the‐art techniques for downscaling climate projections. They are being applied under Perfect‐Prognosis (trained in a historical period with observations) and hybrid approaches (as Regional Climate Models (RCMs) emulators), with satisfactory results. Nevertheless, two important aspects have not been, to our knowledge, properly assessed yet: (1) their performance as emulators for other Earth System Models (ESMs) different to the one used for training, and (2) their performance under extrapolation, that is, when applied outside of their calibration range. In this study, we use UNET, a popular CNN, to assess these two aspects through two pseudo‐reality experiments, and we compare it with simpler emulators: an interpolation and a linear regression. The RCA4 regional model, with 0.11° resolution over a complex domain centered in the Pyrenees, and driven by the CNRM‐CM5 global model is used to train the emulators. Two frameworks are followed for the training: predictors are taken (1) from the upscaled RCM and (2) from the ESM. In both frameworks, the performance of the UNET when applied for other ESMs different to the one used for training is considerably worse, indicating poor generalization. For the linear method a similar deterioration is seen, so this limitation does not seem method specific but inherent to the task. For the second experiment, the emulators are trained in present and evaluated in future, under extrapolation. While averaged aspects such as the mean values are well simulated in future, significant biases (up to 5°C) appear when assessing warm extremes. These biases are larger by UNET than those produced by the linear method. This limitation suggests that, for variables such as temperature, with a marked signal of change and a strong linear relationship with predictors, simple linear methods might be more appropriate than the sophisticated deep learning techniques. [ABSTRACT FROM AUTHOR]

Climatic statistical downscaling in arid and topographically complex river basins remains relatively lacking. To address this gap, climatic variables derived from a global climate model (GCM) ensemble were downscaled from a grid resolution of 2.5° × 2.5° down to the station level. For this purpose, a combination of multiple linear and logistic regressions was developed, calibrated and validated with regard to their predictions of monthly precipitation and daily temperature in the Jordan River Basin. Seasonal standardized predictors were selected using a backward stepwise regression. The validated models were used to examine future scenarios based on GCM simulations under two Representative Concentration Pathways (RCP4.5 and RCP8.5) for the period 2006–2050. The results showed a cumulative near-surface air temperature increase of 1.54 °C and 2.11 °C and a cumulative precipitation decrease of 100 mm and 135 mm under the RCP4.5 and RCP8.5, respectively, by 2050. This pattern will inevitably add stress to water resources, increasing management challenges in the semi-arid to arid regions of the basin. Moreover, the current application highlights the potential of adopting regression-based models to downscale GCM predictions and inform future water resources management in poorly monitored arid regions at the river basin scale. [ABSTRACT FROM AUTHOR]

Climate change is a global challenge necessitating adaptation at the local level. Small island developing states (SIDS) in the southwest Indian Ocean (SWIO) basin are particularly vulnerable and already facing significant challenges due to climate variability and extreme weather events like tropical cyclones (TCs). Tailored climate services, catering to their specific needs and contexts, are crucial for formulating appropriate adaptation strategies. This study aims to fill the gap of localized and reliable information for climate services in the SWIO region. A 12-km resolution regional climate model is dynamically downscaled from a CMIP6 model to capture the climatology of tropical cyclones. Outputs are bias-corrected at kilometer-scale resolution over multiple islands. A subset of CMIP6 simulations is also statistically downscaled over La Réunion to quantify model uncertainties at the local scale and compare statistical and dynamical downscaling methods.CMIP6 models project an average increase in daily mean temperatures over small islands ranging from 1.2 °C (SSP1-2.6) to 3.7 °C (SSP5-8.5) by the end of the century compared to the reference period of 1981–2010, and up to 4.4 °C on Madagascar (SSP5-8.5). The dry season in the SWIO region is anticipated to become significantly drier, with precipitation deficits ranging from 10 % to 40 %, primarily due to a delayed onset of the rainy season. The cyclonic risk is expected to increase due to stronger cyclone intensities, a higher proportion of strong TCs, and the poleward migration of very intense TCs by one to two degrees latitude, further amplifying the risk faced by the Mascarene islands.

Damming and flow regulation ensure water availability during dearth periods, but they cause impacts on inhabiting biota and facilitate the establishment of Invasive Alien Species (IAS). Setting environmental flows (e-flows) has proven fundamental to ensure good ecological status of rivers, battle IAS, and sustain ecosystem services. However, ongoing climate change is further straining water resources availability and the fulfilment of e-flows in Spain, which has compelled River Basin Management Plans (RBMPs) to include specific adaptation measures. In this study, we assessed adaptation strategies addressing land use and climate change in the Serpis River (Eastern Spain), based on increasing irrigation efficiency (i.e. moving from gravity to drip irrigation). The evaluation considered two CMIP6 climate change scenarios (Shared Socioeconomic Pathways SSP1-2.6 and SSP5-8.5) for two time horizons (short term 2015–2040 and mid-term 2041–2070), together with three land use change scenarios (changes in irrigated areas and technologies). Future estimations included reservoir flow releases, e-flows fulfilment, agricultural benefits, and changes in the suitable habitat and potential competition for two native (Eastern Iberian chub Squalius valentinus and European eel Anguilla anguilla) and two invasive (common bleak Alburnus alburnus and pumpkinseed Lepomis gibbosus) fish species. Predictions indicated a decrease in the suitable habitat for eel and Eastern Iberian chub, but the established e-flows will be ensured. Although the maximum benefits will rarely be achieved, on average the modernisation expected in the RBMP would sustain agricultural incomes in the optimistic scenario (SSP1-2.6) and for the short term of the pessimistic scenario (SSP5-8.5), showing similar effects than the ongoing redistribution of agricultural plots. Habitat and competition with invasive species will experience little changes due to the strong flow regulation, although the invasive bleak would prevail in the mid-term of the most pessimistic scenario. These results highlight the necessity of proactive adaptation measures beyond those described in the RBMPs (e.g. forest management or crop substitution) to increase water availability and sustain agricultural benefits, in particular for the most pessimistic scenario. They also underscore the difficulties to counteract invasive species through streamflow management under water scarcity scenarios.

Statistical downscaling (SD) is preferable to dynamic downscaling to derive local-scale climate change information from large-scale datasets. Many statistical downscaling models are available these days, but comparison of their performance is still inadequately addressed for choosing a reliable SD model. Thus, it is desirable to compare the performance of SD models to ensure their adaptability in future climate studies. In this study, a statistical downscaling model (SDSM) or multi-linear regression and the Least Square Support Vector Machine (LS-SVM) were used to do downscaling and compare the results with those obtained from general circulation model (GCM) for identifying the best SD model for the Indira Sagar Canal Command area located in Madhya Pradesh, India. The GCM, Hadley Centre Coupled Model version 3 (HadCM3), was utilized to extract and downscale precipitation, maximum temperature (Tmax), and minimum temperature (Tmin) for 1961–2001 and then for 2001–2099. Before future projections, both SD models were initially calibrated (1961–1990) and validated (1991–2001) to evaluate their performance for precipitation and temperature variables at all gauge stations, namely Barwani, East Nimar, and West Nimar. Results showed that the precipitation trend was under-predicted owing to large errors in downscaling, while temperature was over-predicted by SD models. HIGHLIGHTS Precipitation values are under-predicted, while temperature values are over-predicted by statistical downscaling models.; Large errors in downscaling precipitation are observed, since downscaling of precipitation is more problematic than temperature.; Statistical measures (R2, RMSE, SSE, NSE, and MAE) showed good agreement between observed and downscaled climate variables for SDSM and LS-SVM.;

Extensive investigations on the projection of heat waves (HWs) were conducted on the basis of coarse-resolution global climate models (GCMs). However, these investigations still fail to characterise the future changes in HWs regionally over China. PRECIS dynamical downscaling with a horizontal resolution of 25 km × 25 km was employed on the basis of GCM-HadCM3 to provide reliable projections on HWs over the Chinese mainland, and six statistical downscaling methods were used for bias correction under RCP4.5 and RCP8.5 scenarios. The multi-method ensemble (MME) of the top three dynamical downscaling methods with good performance was used to project future changes. Results showed that PRECIS primarily replicated the detailed spatiotemporal pattern of HWs. However, PRECIS overestimated the HWs in the Northwest and Southeast and expanded the areas of HWs in the Northeast and Southwest. Three statistical downscaling methods (quantile mapping, CDF-t and quantile delta mapping) demonstrated good performance in improving PRECIS simulation for reproducing HWs. By contrast, parametric-based trend-preserving approaches such as scaled distribution mapping and ISI-MIP are outperformed by the three aforementioned methods in downscaling HWs, particularly in the high latitudes of China. Based on MME projections, at the end of the 21st century, the national average of the number of HW days each year, the length of the longest HW event in the year and the extreme maximum temperature in HW will increase by 3 times, 1 time and 1.3 °C, respectively, under the RCP4.5 scenario, whilst that under the RCP8.5 scenario will increase by 8 times, 3 times and 3.7 °C, respectively, relative to 1986–2005. The Northwest is regionally projected to suffer long and hot HWs, whilst the South and Southeast will experience frequent consecutive HWs. Thus, HWs projected by the combined dynamical and statistical downscaling method are highly reliable in projecting HWs over China.

Climate change, particularly global warming, is a significant environmental issue that has gained widespread attention in recent decades. This study aimed to complement the model for the future by utilizing Global Climate Models (GCMs) data. The shallow-layered Artificial Neural Network (ANN) and deep-based Long Short-Term Memory (LSTM) network was applied to extract the historical temperature trend of Calgary, Canada. Mutual Information (MI) was employed for screening purposes to ensure the quality of the input variables. The results of the study indicate that the LSTM model, which relied on the data screening method using MI, achieved RMSE of 0.01 °C, DC of 0.93, a CC of 0.75 and a Bias of 1.89, and has superiority over the ANN method in the Alberta region. [ABSTRACT FROM AUTHOR]

Deep learning methods can achieve a finer refinement required for downscaling meteorological elements, but their performance in terms of bias still lags behind physical methods. This paper proposes a statistical downscaling network based on Light-CLDASSD that utilizes a Shuffle–nonlinear-activation-free block (SNBlock) and Swin cross-attention mechanism (SCAM), and is named SNCA-CLDASSD, for the China Meteorological Administration Land Data Assimilation System (CLDAS). This method aims to achieve a more accurate spatial downscaling of a temperature product from 0.05° to 0.01° for the CLDAS. To better utilize the digital elevation model (DEM) for reconstructing the spatial texture of the temperature field, a module named SCAM is introduced, which can activate more input pixels and enable the network to correct and merge the extracted feature maps with DEM information. We chose 90% of the CLDAS temperature data with DEM and station observation data from 2016 to 2020 (excluding 2018) as the training set, 10% as the verification set, and chose the data in 2018 as the test set. We validated the effectiveness of each module through comparative experiments and obtained the best-performing model. Then, we compared it with traditional interpolation methods and state-of-the-art deep learning super-resolution algorithms. We evaluated the experimental results with HRCLDAS, national stations, and regional stations, and the results show that our improved model performs optimally compared to other methods (RMSE of 0.71 °C/0.12 °C/0.72 °C, BIAS of −0.02 °C/0.02 °C/0.002 °C), with the most noticeable improvement in mountainous regions, followed by plains. SNCA-CLDASSDexhibits the most stable performance in intraday hourly bias at temperature under the conditions of improved feature extraction capability in the SNBlock and a better utilization of the DEM by the SCAM. Due to the replacement of the upsampling method from sub pixels to CARAFE, it effectively suppresses the checkerboard effect and shows better robustness than other models. Our approach extends the downscaling model for CLDAS data products and significantly improves performance in this task by enhancing the model's feature extraction and fusion capabilities and improving upsampling methods. It offers a more profound exploration of historical high-resolution temperature estimation and can be migrated to the downscaling of other meteorological elements. [ABSTRACT FROM AUTHOR]

Deep learning (DL) has emerged as a promising tool to downscale climate projections at regional‐to‐local scales from large‐scale atmospheric fields following the perfect‐prognosis approach. Given their complexity, it is crucial to properly evaluate these methods, especially when applied to changing climatic conditions where the ability to extrapolate/generalize is key. In this work, we intercompare several DL models extracted from the literature for the same challenging use‐case (downscaling temperature in the CORDEX North America domain) and expand standard evaluation methods building on eXplainable Artificial Intelligence (XAI) techniques. Specifically, we introduce two novel XAI‐based diagnostics—Aggregated Saliency Map and Saliency Dispersion Maps—and show how they can be used to unravel the internal behavior of these models, aiding in their design and evaluation. This work advocates for the introduction of XAI techniques into deep downscaling evaluation frameworks, especially when working with large regions and/or under climate change conditions. Plain Language Summary: Due to limitations in the computational resources available, General Circulation Models (GCMs) are often used to simulate the climate system over coarse resolution grids. This hampers the applicability of GCM products in the regional‐to‐local scale, highly demanded by different socio‐economic sectors. Statistical downscaling aims to solve this problem by generating high‐resolution climate fields. Recently, machine learning techniques—particularly deep learning (DL) models—have shown promising results in this task. These models are first trained in a historical period through observational data sets, and then applied to the GCM outputs of plausible future scenarios, thus generating high‐resolution climate change products. To assess the performance of these methods, a number of evaluation metrics have been proposed considering both the skill to reproduce present climate conditions and the ability to generalize changing conditions. Here, we illustrate the possibilities of eXplainable Artificial Intelligence (XAI) techniques to expand the evaluation framework for deep downscaling methods, introducing new XAI‐derived diagnostics to unravel their internal behavior. The results show the usefulness of incorporating XAI techniques into statistical downscaling evaluation frameworks, especially when working with large regions and/or under climate change conditions. Key Points: Explainable artificial intelligence (XAI) facilitates the evaluation of deep downscaling models by unraveling their internal behaviorXAI techniques can detect structural problems not revealed by standard evaluation, which may be relevant for understanding model differencesXAI techniques can assess the relevance and locality of the predictors of deep learning models, helping to assess their physical consistency [ABSTRACT FROM AUTHOR]

In this study, hydrological modeling at the watershed level is used to assess the impacts of climate and land use changes on the catchment area of the Khanpur Dam, which is an important water source for Rawalpindi and Islamabad. The hydrological impact of past and anticipated precipitation in the Khanpur Dam watershed was forecast by using a HEC-HMS model. After calibration, the framework was employed to analyze the effects of changes in land cover and climate on the hydrological regime. The model used information from three climatic gauge stations (Murree, Islamabad Zero Point, and Khanpur Dam) to split the Khanpur Dam catchment area into five sub-basins that encompass the entire watershed region, each with distinctive characteristics. The model was evaluated and checked for 2016–2018 and 2019–2020, and it produced an excellent match with the actual and anticipated flows. After statistical downscaling with the CMhyd model, the most effective performing GCM (MPI-ESM1-2-HR) among the four GCMs was chosen and used to forecast projections of temperature and precipitation within two shared socioeconomic pathways (SSP2 and SSP5). The predictions and anticipated changes in land cover were incorporated into the calibrated HEC-HMS model to evaluate the potential impact of climate change and land cover change at the Khanpur Dam. The starting point era (1990–2015) and the projected period (2016–2100), which encompassed the basis in the present century, were analyzed annually. The results indicated a spike in precipitation for the two SSPs, which was predicted to boost inflows all year. Until the end of the twenty-first century, SSP2 predicted a 21 percent rise in precipitation in the Khanpur Dam catchment area, while SSP5 predicted a 28% rise in precipitation. Increased flows were found to be projected in the future. It was found that the calibrated model could also be used effectively for upcoming studies on hydrological effects on inflows of the Khanpur Dam basin. [ABSTRACT FROM AUTHOR]

Climate model (CM) statistical downscaling requires quality and quantity modifications of the CM's outputs to increase further modeling accuracy. In this respect, multi-resolution wavelet transform (WT) was employed to determine the hidden resolutions of climate signals and eliminate bias in a CM. The results revealed that the newly developed discrete wavelet transform (DWT)-based bias correction method can outperform the quantile mapping (QM) method. In this study, wavelet coherence analysis was utilized to assess the high common powers and the multi-scale correlation between the predictors and predictand as a function of time and frequency. Thereafter, to rate the most contributing predictors based on potential periodicity, the average variance was calculated, which is named the Scaled Average (SA) measure. Consequently, WT along with Artificial Neural Network (ANN) were applied for bias correction and identifying the dominant predictors for statistical downscaling. The CAN-ESM5 data of Canadian climate models and INM-CM5 data of Russian climate models over two climatic areas of Iran with semi-arid (Tabriz) and humid (Rasht) weather were applied. The projection of future precipitation revealed that Tabriz will experience a 3.4–6.1% decrease in precipitation, while Rasht's precipitation will decrease by 1.5–2.5%. These findings underscore the importance of refining CM data and employing advanced techniques to assess the potential impacts of climate change on regional precipitation patterns. [ABSTRACT FROM AUTHOR]

This study evaluates statistical downscaling techniques using different metrics and compares climate change signals and extreme precipitation and temperature changes under future climate change scenarios in the Bosque watershed, North-Central Texas. The study utilizes observed gridded Daymet data to assess the effectiveness of statistical downscaling techniques. It involves comparing the mean, the 90th percentile, 10th percentile, wet day frequency, and Cumulative Distribution Function (CDF) of climate model simulations before and after downscaling and the Daymet data during the historical period (1981–2005). Furthermore, the study analyzes changes in climate change signals, extreme precipitation, and temperature values under both near-future (2031–2060) and far-future (2070–2099) climate scenarios. The Ratio Delta method (DeltaSD) and Equi-Distant Quantile Mapping (EDQM) statistical downscaling techniques adjust the mean annual, the wet days frequency, the 90th and 10th percentiles, and the CDF of Global Climate Models (GCMs) simulations of historical precipitation and temperature. The downscaling techniques influenced the climate change signal and changes in extreme values in the future climate. When examining future climate projections produced using the DeltaSD method, we observe a more pronounced reduction in precipitation, while simulations generated through EDQM exhibit a higher frequency of heavy precipitation events (R10mm, R20mm) and consecutive dry days (CDD). It's worth noting that the uncertainties associated with the statistical downscaling techniques are relatively small and not statistically significant (≤0.05). In contrast, substantial and significant uncertainties arise from the choice of emission scenarios and the selection of driving GCMs. Across most climate change scenarios, there is a consistent trend towards increased temperatures and extreme temperature indices. The trend of extreme temperature indices shows variation following the choice of emission scenarios where a significant change in temperature extremes was detected under the RCP8.5 emission scenario.

The objective of this research was to develop a statistical downscaling approach in the Phetchaburi River Basin, Thailand, consisting of two main processes: predictor selection and relationship construction between predictors and local rainfall. Super predictor (SP) and stepwise regression (SR) were employed with principal component analysis (PCA) and the statistical downscaling model (SDSM) was applied later. Four statistical models were finally used: M1 (SP–PCA), M2 (SP–SDSM), M3 (SR–PCA), and M4 (SR–SDSM) with 26 large circulation indices generated by CanESM2 (CMIP5). Finally, the characteristics of extreme rainfall events were observed under three climate change scenarios during three different periods (2020–2040, 2041–2070, and 2071–2100). The results revealed that rainfall and geostrophic airflow velocity were the best predictors for rainfall downscaling, followed by divergence, meridional velocity, and relative humidity. Three objective functions (R2, NSE, and RMSE) were applied to evaluate model performance. The M4 model presented the highest performance while M1 showed the lowest skill. The average annual rainfall specifically increased compared with the historical rainfall for RCP2.6, RCP4.5, and RCP8.5 scenarios in future periods. Very to extremely wet years determined by the standardized anomaly index (SAI) occurred more often in the far-future while severe to extremely dry years frequently occurred in the mid- and far-future. HIGHLIGHTS The use of stepwise regression and SDSM presented the highest performance for rainfall downscaling.; The two most predictive variables generated by GCMs for rainfall downscaling in this area were precipitation and geostrophic airflow velocity at 850 hPa.; Future rainfall projection trends and their characteristics using the SAI method were presented under three climate change scenarios.;

This study investigates the future changes in precipitation extreme indices in the Lake Urmia Basin during the period 2021–2100 compared to the base period (1987–2016), using the Coupled Model Intercomparison Project Phase 5 models. Trend analysis was performed using Mann–Kendall test and Sen's estimator. The output of these models was downscaled by the Long Ashton Research Station Weather Generator method for the representative concentration pathway (RCPs) 4.5 and 8.5. A model averaging technique was employed to create an ensemble model. The results showed that the average precipitation of the basin will decrease by the end of the 21st century. The projection also showed that the consecutive dry day's index increases based on both scenarios. However, other indices (maximum 1-day precipitation, maximum 5-day precipitation, very wet days, consecutive wet days, simple daily intensity index, and wet-day precipitation) are reduced compared to the base period. Moreover, the slope of significant trends in the RCP8.5 is greater and more severe than that in RCP4.5. HIGHLIGHTS This study focuses on the future change of extreme rainfall in the Lake Urmia basin in Iran.; The projected results are decreasing trends in extreme precipitation and in wet days and increasing trends in dry days by the end of the 21st century.; This study predicts that the basin may face more drought in the future.; The results of this study have important implications for the management of water resources.;

We analyzed hydrological responses to changing climate and land use/land cover (LULC) for the past (1985–2020) and future (2021–2080) in the Chemoga watershed of the Upper Blue Nile Basin. The watershed comprises four agroecological environments: Moist Kolla, Moist Weyna Dega, Moist Dega, and Wet Wurch. Past and projected LULC changes under business-as-usual (BAU) and land conservation (LC) scenarios were utilized. Climate projections from 2021 to 2080, under two Shared Socioeconomic Pathways (SSP2-4.5 and SSP5-8.5), were downscaled from Global Climate Models. Utilizing the Soil and Water Analysis Tool, we assessed impacts on mean annual surface runoff (SR) and evapotranspiration (ET). Maximum and minimum temperatures increased significantly in the past and future climate scenarios, with a significant rainfall increase observed under SSP5-8.5. Historical trends revealed a 16.6% increase in SR and 7% in ET from 1983–2002 to 2003–2020. Under BAU LULC with the SSP2-4.5 (SSP5-8.5) climate scenario, SR increased by 24% (26.1%) and ET by 3.1% (4.4%) from 2003–2020 to 2021–2050, followed by a subsequent SR rise of 13.7% (14.0%) and ET increase of 6.0% (5.7%) from 2021–2050 to 2051–2080. Conversely, the LC LULC with SSP2-4.5 (SSP5-8.5) resulted in a 5.3% (4.2%) SR decrease and ET increase of 9.7% (11.3%) from 2003–2020 to 2021–2050 and a further SR decrease of 1% (0.7%) and 6.1% (6.9%) ET increase from 2021–2050 to 2051–2080. The Moist Kolla agroecology experienced the highest SR increase due to vegetation clearances for commercial farming. Meanwhile, the LC scenario indicated substantial decreases in SR and marginal increases in ET in the Moist Weyna Dega agroecology due to forest restoration on steep slopes. Overall, SR showed greater sensitivity to LULC changes, while ET was more responsive to climate changes. The results emphasize considering diverse agroecological contexts for effective water resource management under changing climate and LULC scenarios.

This study aims to evaluate the impact of climate change on the surface water hydrology of the Gopad river basin in India. The outputs of four CMIP6 Global Climate Models have been downscaled using the statistical downscaling method to the basin level. A comparative analysis for the accuracy achieved in the bias correction for the combination of GCM and downscaling method has been performed before utilising the downscaled weather parameters for hydrological study. The MIROC6 and ACCESS-CM2 were found best for the simulation of precipitation and temperature, respectively. The Distribution Mapping and Variance Scaling methods have shown better accuracy w.r.t other statistical methods. The impact of climate change has been found significant since the temperature has been observed to be increased by 3.16 °C by the end of 2060; meanwhile, there is an average decrease of 9.2% in the annual rainfall from the baseline. The peak runoff has increased while there is a significant decrease in the groundwater recharge. Further, hydrologically critical subbasins (HCS) have been delineated based on the runoff, groundwater recharge, and baseflow. Most HCS was observed to be situated in the upper Gopad river basin, representing the area's pristine conditions. HIGHLIGHTS The ACCESS-CM2 and MIROC6 are more accurate for the study area.; The Variance Scaling and Distribution Mapping methods have shown better accuracy than others.; A rise in average temperature has been observed, while there is a decrease in annual rainfall from the baseline.; The area with less human intervention has proven to be hydrologically critical for the Gopad river basin.;

The climate change impact and adaptation communities need future scenarios with sufficient high resolution, which are frequently achieved by applying Statistical Downscaling Models (SDMs) over global climate models. A large variety of SDMs exists, and some can be more suitable than others for each specific purpose. For this reason, it is important to develop tools to facilitate the evaluation and generation of downscaled scenarios following different approaches. In this paper we present a service, ‘pyClim-SDM’, which allows users to generate and evaluate their own downscaled scenarios with a very simple and user-friendly graphical interface. This tool includes a large collection of state-of-the-art methods belonging to different families to downscale daily data of the following surface variables: temperature, precipitation, wind, relative humidity and cloud coverage. Additionally, the software is prepared to be applied over any other user-defined target variable. Thus, multivariable indexes can be tackled as target variables themselves, instead of being calculated from the downscaled primary variables. With this possibility, potential intervariable inconsistencies are avoided. An application example for a Fire Weather Index, dependent on temperature, wind, humidity and precipitation, is shown. The service here presented -mainly based on a new downscaling software and a user-friendly graphical interface- is an essential piece for evaluating and generating high-resolution projection data within the Spanish national climate change adaptation strategy which includes, among other elements, a common database for all sectors, viewer and data distribution portal, etc.

Abstract Deep learning (DL) has emerged as a promising tool to downscale climate projections at regional‐to‐local scales from large‐scale atmospheric fields following the perfect‐prognosis approach. Given their complexity, it is crucial to properly evaluate these methods, especially when applied to changing climatic conditions where the ability to extrapolate/generalize is key. In this work, we intercompare several DL models extracted from the literature for the same challenging use‐case (downscaling temperature in the CORDEX North America domain) and expand standard evaluation methods building on eXplainable Artificial Intelligence (XAI) techniques. Specifically, we introduce two novel XAI‐based diagnostics—Aggregated Saliency Map and Saliency Dispersion Maps—and show how they can be used to unravel the internal behavior of these models, aiding in their design and evaluation. This work advocates for the introduction of XAI techniques into deep downscaling evaluation frameworks, especially when working with large regions and/or under climate change conditions.

This study analyzed the main features (onset, demise, and length) of the South American Monsoon System (SAMS) projected in different time slices (2020–2039, 2040–2059, 2060–2079, and 2080–2099) and climate scenarios (SSP2–4.5 and SSP5–8.5). Eight global climate models (GCMs) from the Coupled Model Intercomparison Project Phase 6 (CMIP6) that perform well in representing South America's historical climate (1995–2014) were initially selected. Thus, the bias correction–statistical downscaling (BCSD) technique, using quantile delta mapping (QDM), was applied in each model to obtain higher-resolution projections than their original grid. The horizontal resolution adopted was 0.5° of latitude × longitude, the same as the Climate Prediction Center precipitation analysis used as a reference dataset in BCSD. The QDM technique improved the monsoon onset west of 60° W and the simulated demise and length in southwestern Amazonia. Raw and BCSD ensembles project an onset delay of approximately three pentads compared to the historical period over almost all regions and a demise delay of two pentads northward 20° S. Additionally, the BCSD ensemble projects a reduced length with statistical significance in most South Atlantic Convergence Zone regions and a delay of three pentads in the demise over the Brazilian Amazon from the second half of the 21st century. [ABSTRACT FROM AUTHOR]

This paper evaluated and compared the performance of a statistical downscaling method and a dynamical downscaling method to simulate the spatial–temporal rainfall distribution. Outputs from RegCM4 Regional Climate Model (RCM) and the CanESM2 Atmosphere–Ocean General Circulation Model (AOGCM) were selected for the data scarce Gwadar‐Ormara basin, Pakistan. The evaluation was based on the climatological average and standard deviation for historic (1971–2000) and future (2041–2070) time periods under Representative Concentration Pathways (RCP) 4.5 and 8.5 scenarios. The performance evaluation showed that statistical downscaling is preferred to simulate and project rainfall patterns in the study area. Additionally, the Statistical DownScaling Model (SDSM) showed low R2 values in calibration and validation of the simulations with respect to observed data for the historic period. Overall, SDSM generated satisfactory results in simulating the monthly rainfall cycle of the entire basin. In this study, RegCM4 showed large rainfall errors and missed one rainfall season in the historic period. This study also explored whether the grid‐based rainfall time series of the Asian Precipitation—Highly Resolved Observational Daily Integration Towards Evaluation (APHRODITE) dataset could be used to enlarge and complement the sample of in situ observed rainfall time series. A spatial correlogram was used for observed and APHRODITE rainfall data to assess the consistency between the two data sources, which resulted in rejecting APHRODITE data. For the future time period (2041–2070) under RCPs 4.5 and 8.5 scenarios, rainfall projections did not show significant difference for both downscaling approaches. This may relate to the driving model (CanESM2 AOGCM) and not necessarily suggests poor performance of downscaling; either statistical or dynamical. Hence, the study recommends evaluating a multi‐model ensemble including other GCMs and RCMs for the same area of study. [ABSTRACT FROM AUTHOR]

Introduction and Objective: The problem of drought and water crisis in the in the central Iran watershed has become serious in recent years due to consecutive droughts and increased water consumption in drinking, agriculture and industry sectors. Since rain is the only source of water supply in this catchment area, a decrease in the amount of rainfall with this intensity can cause a widespread drought and eventually large migrations from these areas. Therefore, the aim of this research is to analyze the changes in the aridity index in the vast area of the central Iran watershed in future. Material and Methods: In order to achieve the study's objective, the minimum and maximum air temperature, sunshine hours, and precipitation of 40 synoptic stations in the Central Iran watershed for a period of 20 years (1991-2010) were received from the Meteorological Organization.Then the values of potential evapotranspiration, aridity index, and Angot precipitation index were calculated for each station in the current conditions. In order to project changes in the aridity index in the future, the output of three MIROC5, HadGEM2-ES, and GFDL-CM3 models for the historical period (1991–2010) and the medium future (2041–2060) under the RCP4.5 scenario and based on the LARS- WG6 scaled down. After evaluating the accuracy of the models against the observational data, the index values were calculated in future conditions, and zoning maps were drawn for the study area. Results: Projection of changes in evapotranspiration values under the RCP4.5 scenario (2041– 2060) showed that the most changes were in January, March, December, and November. Shahrekord showed the highest percentage of incremental changes, with 36.8% change in January compared to the base period. In terms of rainfall regime, the Central Iran watershed has a winter rainfall regime based on the Angot precipitation index in the current conditions. The HadGEM2-ES model is consistent with the base period, but based on the MIROC5 and GFDLCM3 models, the winter precipitation in many stations will change to the all-season precipitation regime in the future. Climatic zoning of the UNEP aridity index in the studied area indicated that in current conditions, more than half of the stations have a dry climate. Only Kohrang has a humid climate with an aridity index value of 0.78. Shemiranat and Oligoders have semi-dry conditions with values of 0.21 and 0.20, respectively. The evaluation of the future climate conditions according to the GFDL-CM3 model shows that the humid climate in central Iran will gradually disappear, while the semi-arid climate will expand in this region. The output of the MIROC5 model also shows the expansion of the ultra-arid climate zone in different parts of the catchment area of Central Iran. Conclusion: The results showed that the aridity index values based on selected GCM models have decreased in the studied area compared to the observation period. Therefore, a large number of stations that have dry conditions in the base period will experience ultra-dry conditions in the future climate.The evapotranspiration potential in the Central Iran watershed is increasing during all months of the year, especially in wet stations in the west. This issue shows the dominance of the aridity index over this region in future conditions. [ABSTRACT FROM AUTHOR]

Drought events are critical environmental threats that yield several socioeconomic impacts. Such effects are even more relevant for South America (SA) since different activities essential for the continent, such as agriculture and energy generation, depend highly on water resources. Thus, this study aimed to evaluate future changes in precipitation and hydrological drought occurrence in SA through climate projections from eight global climate models (GCMs) of CMIP6. To this end, statistical downscaling was applied to the projections obtained using the quantile delta mapping technique, and the method proved to be efficient in reducing systematic biases and preserving GCMs' trends. For the following decades, the results show considerable and statistically significant reductions in precipitation over most of SA, especially during the austral spring, with the most intense signal under the SSP5-8.5 forcing scenario. Furthermore, GCMs showed mixed signals about projections of the frequency and intensity of drought events. Still, they indicated agreement regarding the increased duration and severity of events over the continent and a substantial proportion of moderate and severe events over most of Brazil during the 21st century. These results can be helpful for better management of water resources by decision-makers and energy planners. [ABSTRACT FROM AUTHOR]

High spatio‐temporal resolution near‐surface projected data is vital for climate change impact studies and adaptation. We derived the highest statistically downscaled resolution multivariate ensemble currently available: daily 1 km until the end of the century. Deep learning models were employed to develop transfer functions for precipitation, water vapor pressure, radiation, wind speed, and, maximum, mean and minimum temperature. Perfect prognosis is the particular statistical downscaling methodology applied, using a subset of the ReKIS data set for Saxony as predictands, the ERA5 reanalysis as during‐training predictors and the CORDEX‐EUR11 ensemble as projected predictors. The performance of the transfer functions was validated with the VALUE framework, yielding highly satisfactory results. Particular attention was given to the three major perfect prognosis assumptions, for which several tests were carried out and thoroughly discussed. From the latter, we corroborated their fulfillment to a high degree, thus, the derived projections are considered adequate and relevant for impact modelers. In total, 18 runs for RCP85, 1 for RCP45, and 4 for RCP26 were downscaled under both stochastic and deterministic approaches. This multivariate ensemble could drive more accurate and diverse impact studies in the region. Generally, the projected climatologies are in agreement with coarser resolution projections. Nevertheless, statistical particularities were observed for some projections, thus, a list of caveats for potential users is given. Due to the scalability of the presented methodology, further possible applications with additional datasets are proposed. Lastly, several potential improvement prospects are discussed toward the ideal subsequent iteration of the perfect prognosis statistical downscaling methodology. Plain Language Summary: There is a great worldwide demand for high spatio‐temporal resolution projections to develop climate change adaptation and mitigation schemes. Despite recent improvements, the resolution of both global and regional climate models is still too coarse to properly represent local variability, particularly in complex terrains. Depending on the application, impact modelers and decision makers require kilometer‐scale projections, with a minimum daily temporal resolution, of near‐surface variables. To fill this information gap, we employed artificial intelligence algorithms to downscale, to a novel daily 1 km resolution, a projection ensemble until the end of the century consisting of precipitation, water vapor pressure, radiation, wind speed, and, maximum, mean and minimum temperature. The ensemble comprises 18 runs of the business‐as‐usual worst‐case scenario (RCP85), 1 run of the stabilization scenario (RCP45), and 4 of the optimistic low‐emissions scenario (RCP26). The main assumptions of the methodology were thoroughly tested and discussed. The validation carried out yielded highly satisfactory results. Thus, we consider the projections to be adequate and relevant for impact studies. The region studied is located in Saxony (Germany), still, the methodology shown is potentially applicable anywhere in the world. Key Points: Highest statistically downscaled spatio‐temporal resolution multivariate ensemble currently available, consisting of 23 projection runsWe downscaled precipitation, water vapor pressure, radiation, wind speed, and, maximum, mean and minimum temperatureThe methodology complied to a high degree with the three perfect prognosis assumptions and is scalable to other spatio‐temporal resolutions [ABSTRACT FROM AUTHOR]

Study region: Quito, Ecuador Study focus: The study region faces two water-related challenges, which, to date, have only been studied to a minimal extent: extreme precipitation events and water shortage in the dry season. This study investigates the current conditions and future changes in short-duration events, low river discharges and associated uncertainties. Daily precipitation and temperature projections from 19 state-of-the-art global climate models (CMIP6) and four future scenarios (SSP1–2.6, SSP2–4.5, SSP3–7.0 and SSP5–8.5) are downscaled with delta change and more advanced quantile perturbation methodologies. Temporal disaggregation is applied to obtain sub-daily precipitation with five extreme value distributions (EVDs). Three conceptual hydrological models are implemented to quantify the impacts on river flows. Variance decomposition is applied to estimate the uncertainty share of climate models, hydrological models, and EVDs in the results. New hydrological insights: The results indicate an intensification of extreme precipitation events, with short-duration events expected to be more intense by up to 30% in the near future (2021–2050) and up to 60% in the far future (2070–2099). The river peak discharges are projected to increase by 5–20% and 10–50% in the near and far future, respectively. On the contrary, the low river flows are projected to decrease by 0–13% and 0–30% for the respective time horizons. Overall, climate models are the dominant source of uncertainty.

Central Asia (CA) is highly sensitive and vulnerable to changes in precipitation due to global warming, so the projection of precipitation extremes is essential for local climate risk assessment. However, global and regional climate models often fail to reproduce the observed daily precipitation distribution and hence extremes, especially in areas with complex terrain. In this study, we proposed a statistical downscaling (SD) model based on quantile delta mapping to assess and project eight precipitation indices at 73 meteorological stations across CA driven by ERA5 reanalysis data and simulations of 10 global climate models (GCMs) for present and future (2081–2100) periods under two shared socioeconomic pathways (SSP245 and SSP585). The reanalysis data and raw GCM outputs clearly underestimate mean precipitation intensity (SDII) and maximum 1-day precipitation (RX1DAY) and overestimate the number of wet days (R1MM) and maximum consecutive wet days (CWD) at stations across CA. However, the SD model effectively reduces the biases and RMSEs of the modeled precipitation indices compared to the observations. Also it effectively adjusts the distributional biases in the downscaled daily precipitation and indices at the stations across CA. In addition, it is skilled in capturing the spatial patterns of the observed precipitation indices. Obviously, SDII and RX1DAY are improved by the SD model, especially in the southeastern mountainous area. Under the intermediate scenario (SSP245), our SD multi-model ensemble projections project significant and robust increases in SDII and total extreme precipitation (R95PTOT) of 0.5 mm d−1 and 19.7 mm, respectively, over CA at the end of the 21st century (2081–2100) compared to the present values (1995–2014). More pronounced increases in indices R95PTOT, SDII, number of very wet days (R10MM), and RX1DAY are projected under the higher emission scenario (SSP585), particularly in the mountainous southeastern region. The SD model suggested that SDII and RX1DAY will likely rise more rapidly than those projected by previous model simulations over CA during the period 2081–2100. The SD projection of the possible future changes in precipitation and extremes improves the knowledge base for local risk management and climate change adaptation in CA.

Deep learning techniques, especially convolutional neural networks (CNNs), have dramatically boosted the performance of statistical downscaling. In this study, we propose a CNN-based 2 m air temperature downscaling model named Terrain-Guided Attention Network (TGAN), which aims at rebuilding 2 m air temperature distribution from 0.0625$^{\circ }$ to 0.01$^{\circ }$ over Southwest China. More concretely, TGAN utilizes two upsampling modules to progressively reconstruct the high-resolution temperature data from the low-resolution one. Then, to better recover the spatial detail of the low-resolution temperature data, an attentive-terrain block is proposed to introduce digital terrain model (DEM) information. It aggregates the temperature data and the corresponding-scale DEM information via the attention mechanism in a multiscale manner. Ultimately, the reconstruction module is employed to obtain the high-resolution temperature data. We use the 2019 data for training, and utilize the 2018 data to verify the effectiveness of the proposed TGAN. The experimental results showed that TGAN achieved the lowest root-mean-square error (1.12$\,^\circ$C) when incorporating DEM data by attentive-terrain blocks in a multiscale manner, followed by incorporating DEM data in a multiscale manner (TGAN-land, 1.31$\,^\circ$C) and only incorporating DEM data (SRCNN-land, 1.36$\,^\circ$C). Meanwhile, TGAN showed a competitive performance when compared with several advanced deep-learning-based super-resolution algorithms and reconstructed the texture details of 2 m air temperature fields more clearly. In general, among various deep learning approaches, TGAN achieves better downscaling results for 2 m air temperature reconstruction and provides a practical method and guidance for the back-calculation of high-resolution historical meteorological grid data.

Climatic statistical downscaling in arid and topographically complex river basins remains relatively lacking. To address this gap, climatic variables derived from a global climate model (GCM) ensemble were downscaled from a grid resolution of 2.5° × 2.5° down to the station level. For this purpose, a combination of multiple linear and logistic regressions was developed, calibrated and validated with regard to their predictions of monthly precipitation and daily temperature in the Jordan River Basin. Seasonal standardized predictors were selected using a backward stepwise regression. The validated models were used to examine future scenarios based on GCM simulations under two Representative Concentration Pathways (RCP4.5 and RCP8.5) for the period 2006–2050. The results showed a cumulative near-surface air temperature increase of 1.54 °C and 2.11 °C and a cumulative precipitation decrease of 100 mm and 135 mm under the RCP4.5 and RCP8.5, respectively, by 2050. This pattern will inevitably add stress to water resources, increasing management challenges in the semi-arid to arid regions of the basin. Moreover, the current application highlights the potential of adopting regression-based models to downscale GCM predictions and inform future water resources management in poorly monitored arid regions at the river basin scale.

Abstract This paper evaluated and compared the performance of a statistical downscaling method and a dynamical downscaling method to simulate the spatial–temporal rainfall distribution. Outputs from RegCM4 Regional Climate Model (RCM) and the CanESM2 Atmosphere–Ocean General Circulation Model (AOGCM) were selected for the data scarce Gwadar‐Ormara basin, Pakistan. The evaluation was based on the climatological average and standard deviation for historic (1971–2000) and future (2041–2070) time periods under Representative Concentration Pathways (RCP) 4.5 and 8.5 scenarios. The performance evaluation showed that statistical downscaling is preferred to simulate and project rainfall patterns in the study area. Additionally, the Statistical DownScaling Model (SDSM) showed low R2 values in calibration and validation of the simulations with respect to observed data for the historic period. Overall, SDSM generated satisfactory results in simulating the monthly rainfall cycle of the entire basin. In this study, RegCM4 showed large rainfall errors and missed one rainfall season in the historic period. This study also explored whether the grid‐based rainfall time series of the Asian Precipitation—Highly Resolved Observational Daily Integration Towards Evaluation (APHRODITE) dataset could be used to enlarge and complement the sample of in situ observed rainfall time series. A spatial correlogram was used for observed and APHRODITE rainfall data to assess the consistency between the two data sources, which resulted in rejecting APHRODITE data. For the future time period (2041–2070) under RCPs 4.5 and 8.5 scenarios, rainfall projections did not show significant difference for both downscaling approaches. This may relate to the driving model (CanESM2 AOGCM) and not necessarily suggests poor performance of downscaling; either statistical or dynamical. Hence, the study recommends evaluating a multi‐model ensemble including other GCMs and RCMs for the same area of study.

Abstract High spatio‐temporal resolution near‐surface projected data is vital for climate change impact studies and adaptation. We derived the highest statistically downscaled resolution multivariate ensemble currently available: daily 1 km until the end of the century. Deep learning models were employed to develop transfer functions for precipitation, water vapor pressure, radiation, wind speed, and, maximum, mean and minimum temperature. Perfect prognosis is the particular statistical downscaling methodology applied, using a subset of the ReKIS data set for Saxony as predictands, the ERA5 reanalysis as during‐training predictors and the CORDEX‐EUR11 ensemble as projected predictors. The performance of the transfer functions was validated with the VALUE framework, yielding highly satisfactory results. Particular attention was given to the three major perfect prognosis assumptions, for which several tests were carried out and thoroughly discussed. From the latter, we corroborated their fulfillment to a high degree, thus, the derived projections are considered adequate and relevant for impact modelers. In total, 18 runs for RCP85, 1 for RCP45, and 4 for RCP26 were downscaled under both stochastic and deterministic approaches. This multivariate ensemble could drive more accurate and diverse impact studies in the region. Generally, the projected climatologies are in agreement with coarser resolution projections. Nevertheless, statistical particularities were observed for some projections, thus, a list of caveats for potential users is given. Due to the scalability of the presented methodology, further possible applications with additional datasets are proposed. Lastly, several potential improvement prospects are discussed toward the ideal subsequent iteration of the perfect prognosis statistical downscaling methodology.

Study region: Belgium Study focus: We assessed statistical downscaling (21 methods) skill and assumptions for seven extreme rainfall indicators in three Belgian areas. The ensemble consists of (i) perturbation methods with four members, (ii) quantile mapping with ten members, and (iii) machine learning with seven members. The methods were evaluated by a perfect predictor experiment in which the outputs of a high-resolution climate model are considered pseudo-observations. The ALARO-0 model with 4 km spatial resolution was considered the high-resolution climate model. The high-resolution outputs were compared with the downscaled results of three future scenarios of the 50 km resolution ALARO-0 model. New hydrological insights: The methods’ skill depends on the area. The skill is higher for the Belgian coastal area compared to the inland and hilly areas. Methods which are based on stationary assumptions have a good skill for the coastal area but show a lower skill for the other areas where the temporal variability in meteorological conditions is stronger. The higher this temporal variability, the more prone are the stationary assumptions to deteriorate in time under different climate forcing conditions. The quantile-based perturbation methods are overall the most robust. An ensemble approach is advised, where different downscaling methods are applied and the uncertainty in the downscaling taken into account, especially when a pseudo-future evaluation, as done in this study, is not possible.

In order to assess the effects of climate change and land use change on Rawal Dam, a major supply of water for Rawalpindi and Islamabad, this study uses hydrological modeling at the watershed scale. The HEC-HMS model was used to simulate the hydrological response in the Rawal Dam catchment to historical precipitation. The calibrated model was then used to determine how changes in land use and climate had an impact on reservoir inflows. The model divided the Rawal Dam watershed into six sub-basins, each with unique features, and covered the entire reservoir's catchment area using data from three climatic stations (Murree, Islamabad Zero Point and Rawal Dam). For the time spans of 2003–2005 and 2006–2007, the model was calibrated and verified, respectively. An excellent fit between the observed and predicted flows was provided by the model. The GCM (MPI-ESM1-2-HR) produced estimates of temperature and precipitation under two Shared Socioeconomic Pathways (SSP2 and SSP5) after statistical downscaling with the CMhyd model. To evaluate potential effects of climate change and land use change on Rawal Dam, these projections, along with future circumstances for land use and land cover, were fed to the calibrated model. The analysis was carried out on a seasonal basis over the baseline period (1990–2015) and over future time horizon (2016–2100), which covers the present century. The findings point to a rise in precipitation for both SSPs, which is anticipated to result in an increase in inflows throughout the year. SSP2 projected a 15% increase in precipitation across the Rawal Dam catchment region until the end of the twenty-first century, while SSP5 forecasted a 17% increase. It was determined that higher flows are to be anticipated in the future. The calibrated model can also be utilized successfully for future hydrological impact assessments on the reservoir, it was discovered. [ABSTRACT FROM AUTHOR]

Convolutional neural networks (CNNs) have been widely studied and found to obtain favorable results in statistical downscaling to derive high-resolution climate variables from large-scale coarse general circulation models (GCMs). However, there is a lack of research exploring the predictor selection for CNN modeling. This paper presents an effective and efficient greedy elimination algorithm to address this problem. The algorithm has three main steps: predictor importance attribution, predictor removal, and CNN retraining, which are performed sequentially and iteratively. The importance of individual predictors is measured by a gradient-based importance metric computed by a CNN backpropagation technique, which was initially proposed for CNN interpretation. The algorithm is tested on the CNN-based statistical downscaling of monthly precipitation with 20 candidate predictors and compared with a correlation analysis-based approach. Linear models are implemented as benchmarks. The experiments illustrate that the predictor selection solution can reduce the number of input predictors by more than half, improve the accuracy of both linear and CNN models, and outperform the correlation analysis method. Although the RMSE (root-mean-square error) is reduced by only 0.8%, only 9 out of 20 predictors are used to build the CNN, and the FLOPs (Floating Point Operations) decrease by 20.4%. The results imply that the algorithm can find subset predictors that correlate more to the monthly precipitation of the target area and seasons in a nonlinear way. It is worth mentioning that the algorithm is compatible with other CNN models with stacked variables as input and has the potential for nonlinear correlation predictor selection. [ABSTRACT FROM AUTHOR]

Under the perfect prognosis approach, statistical downscaling methods learn the relationships between large‐scale variables from reanalysis and local observational records. These relationships are subsequently applied to downscale future global climate model (GCM) simulations in order to obtain projections for the local region and variables of interest. However, the capability of such methods to produce future climate change signals consistent with those from the GCM, often referred to as transferability, is an important issue that remains to be carefully analyzed. Using the EC‐Earth GCM and focusing on precipitation, we assess the transferability of generalized linear models, convolutional neural networks and a posteriori random forests (APRFs). We conclude that APRFs present the best overall performance for the historical period, and future local climate change signals consistent with those projected by EC‐Earth. Moreover, we show how a slight modification of APRFs can greatly improve the temporal consistency of the downscaled series. Plain Language Summary: Even though they are the main tool to study climate change, global climate models (GCMs) still have a limited spatial resolution (around a hundred kilometers) and exhibit considerable biases with respect to the observed climate. Statistical downscaling aims to solve this issue by learning statistical relationships between large‐scale variables, well reproduced by GCMs (e.g., synoptic winds or specific humidity), and local observations of the target surface variable (e.g., precipitation). These relationships are learned over a historical period, and thus a relevant question is whether they can be transferred to the future GCM simulations, that is, whether climate changes produced by GCMs (e.g., changes in mean rainfall) are broadly preserved by the downscaling methods. The rationale behind this is that, even though GCM simulations are biased, GCMs resolve the physical processes responsible for the evolution of the climate system and these changes are thus physically driven. Using the EC‐Earth GCM, we assess the transferability of three statistical downscaling methods (generalized linear models, convolutional neural networks and a posteriori random forests (APRFs)) for precipitation downscaling over Europe. We intercompare them using several diagnostic metrics, concluding that APRFs produce reliable projections, with future climate changes consistent with those projected by EC‐Earth. Key Points: We assess generalized linear models, convolutional neural networks and a posteriori random forests (APRFs) to downscale precipitation from EC‐EarthThe three methodologies lead to plausible local climate change signals, broadly compatible with those given by the raw outputs from EC‐EarthMoreover, we conclude that APRFs provide the best overall performance, in terms of several metrics [ABSTRACT FROM AUTHOR]

In general, weather forecasting has been significantly developed at a large scale and, joined with statistical techniques, is used to predict at a local scale. However, there is no way to propagate winds between two nearby locations; this is a spatial transference, for example, for the waves. After studying coastal dunar systems affected by winds, we have proposed a way for the spatial propagation of wind for scales under 10 km. The proposed transference is based on local data, and it is developed in an easy and accurate way by different regression methods and the wind profile theory. The aim of this article is to establish a methodology for achieving a wind transfer function for local applications. For this purpose, we analyzed and compared data from a field experiment and from a nearby weather station. A combination of the wind profile and statistical downscaling technique formed the basis of this research, which leads to transfer equations for wind speeds and directions. To clarify the procedure, the proposed methodology was applied to the Valdevaqueros Coastal Dune in order to develop a transfer function using time series data from a nearby meteorological station located in Tarifa. [ABSTRACT FROM AUTHOR]

Land cover change (LCC) and climate change (CC) impacts on streamflow in high elevated catchments are a great challenge to sustainable management and the development of water resources. This study evaluates the possible future impacts of both land cover and climate change on the streamflows in the Mohmand Dam catchment, Pakistan, by utilizing the semi-distributed hydrological model known as the Soil and Water Assessment Tool (SWAT), along with the latest Coupled Model Intercomparison Project phase 6 (CMIP6) dataset of different global climate models (GCMs). The downscaling of the precipitation and temperature data was performed by the CMhyd software. The downscaled precipitation and temperature projections from the best performing GCM, out of four GCMs, under two shared socioeconomic pathways (SSP2 and SSP5) and future land cover conditions were forced in a calibrated hydrological model (SWAT model). Compared to the baseline period (1990–2015), the outputs from the selected GCM indicated an increase in the average monthly precipitation, and the maximum and minimum temperature in the study area under both the SSP2 and SSP5 scenarios, by the end of the 21st century. It is expected that the increase in precipitation for the period 2016–2100 is 10.5% and 11.4% under the SSP2 and SSP5 scenarios, respectively. Simulated results from the SWAT model showed significant impacts from the projected climate and land cover changes on Mohmand Dam flows that include: (a) an increase in the overall mean annual flow ranging from 13.7% to 34.8%, whereas the mean monthly flows of June, July and August decreased, and (b) a shift in the peak flows in the Mohmand catchment from July to June. It is concluded that the projected climate changes can substantially influence the seasonality of flows at the Mohmand Dam site. Climate and land cover change impacts are significant, so project planners and managers must include CC and LCC impacts in the proposed operational strategy. [ABSTRACT FROM AUTHOR]

Global climate change and rapid urbanization increase the risk of urban flooding, especially in China. Climate change and the ‘heat island effect’ have increased the frequency of extreme precipitation. Affected by the backwardness of drainage facilities and the lack of drainage capacity, many cities have experienced large-scale waterlogging in low-lying areas, and ocean-like phenomena appear in cities. The public infrastructure was damaged and caused a lot of economic losses. Therefore, it is important to investigate the adaptability of drainage systems to the future in a changing environment. The Sixth International Coupled Model Intercomparison Project (CMIP6) and Storm Water Management Model (SWMM) were used to quantify the impact of climate change on Beijing's waterlogging under different rainstorm scenarios for the future 40 years. The quantile delta mapping method of daily precipitation based on frequency (DFQDM) is proposed to correct the daily precipitation of the climate model and which is proved to be feasible. After the annual precipitation and extreme precipitation index are corrected, percent bias (PBIAS) is significantly reduced. The PBIAS of the extreme precipitation index of the corrected model is all controlled within 6%. The corrected accuracy of CanESM5 is the best. The total flood volume (TFV) of the node increases with the aggravation of climate change. The TFV of SSP5-8.5 and SSP2-4.5 increased by 45.43 and 20.8% in the 100-year return period, respectively, and more than 94% of the conduits reached the maximum drainage capacity in different return periods. After the low impact development (LID) was installed, the improvement effect on the outflow with a smaller return period was significant, decreasing by about 50%. The LID can effectively reduce the overflow of the drainage system. The results of this study can provide suggestions for the reconstruction of the drainage system and the management of flood risk for Beijing in the future. HIGHLIGHTS The DFQDM method for daily precipitation correction of the climate model is proposed.; The TFV of SSP5-8.5 and SSP2-4.5 increased by 45.43 and 20.8% in the 100-year return period, respectively.; The LID can effectively reduce the overflow of the drainage system.;

Due to climatological changes, a study was conducted in the Gediz Basin, Turkey, where agricultural production holds an important place. In the study prepared, 12 general circulation models (GCMs) were utilized under representative concentration pathway (RCP)4.5, RCP6.0, and RCP8.5 scenarios of the fifth assessment report (AR5) of IPCC for the period 2015–2050. The statistical downscaling methods were employed and the projections were derived right after applying the weighted-averaged ensemble mean by the Bayesian Model Averaging (BMA) method and bias correction by equidistant quantile mapping (EDQM). The temperature-based potential evapotranspiration (PET) formulas were modified in accordance with the Penman–Monteith method and the aridity indexes were calculated by UNEP's formula. According to the projections, the mean annual temperature increases between 1.5 and 2.2 °C and the mean total annual PET increases between 5 and 8% are foreseen in the Gediz Basin for the near future. It is foreseen that a semi-arid climate regime may predominate over the region for all of the RCP scenarios under the increasing dryness in basin climate. In addition, it was obtained in the study that sub-humid climate state occurrence for all of the regions included by the basin may be unexpected in the future for the RCP8.5 scenario. The presence of semi-arid climate conditions may be more potent with the increasing trend of radiative forcing over time. HIGHLIGHTS Artificial intelligence methods were performed for downscaling.; A multi-model ensemble strategy was carried out for climate projection.; A robust bias-correction method was utilized.; Climate change forecasts were integrated with aridity concepts.; Spatio-temporal aridity changes were presented for the future term.;

Natural disasters related to water resources as a result of climate change are becoming a major concern for the entire world. The majority of people in India work in agriculture. Droughts, floods, cyclones, and other natural disasters are all linked to uncertainty in meteorological parameters. The primary goal of this research is to investigate and forecast various meteorological parameters in the Bhavnagar district of the Gujarat region. The statistical downscaling technique uses Global Climate Model (GCM) data for analysis, while the dynamic downscaling technique uses Regional Climate Model (RCM) data. RCM data are prepared for important places of the Earth. GCM data are freely available for all parts of the Earth. So, in this study, GCM data are used. RCM data and dynamic downscaling make the process complicated. Using GCM data and the statistical downscaling technique, meteorological parameters i.e. maximum temperature, minimum temperature, and rainfall can be accurately predicted in this study. To predict meteorological parameters, RCP 8.5 scenario and SPI are used to identify the various types of drought years of the study area. Flood hazard maps and drought contingency plans can be prepared based on predicted meteorological parameters. According to SPI analysis, Bhavnagar will experience 26 mild, 8 moderate, 2 severe, and 3 special droughts up to the year 2100. HIGHLIGHTS Forecasting the precipitation and temperature for Bhavnagar district.; Finding out the characteristic of meteorological parameters of Bhavnagar.; Preparation of model to predict drought years using multi-linear regression.; Identification of drought years up to the year 2100 by statistical downscaling technique.; Classifying predicted drought years according to its severity to guide policy makers.;

General circulation models suggest that changes in climatic parameters will have an effect on food production globally. Therefore, this study assessed the impact of climate change on wheat–maize rotation areas in the Huang–Huai–Hai Plain (3H Plain). The projections generated suggest an increase in precipitation during the wheat growth period (WGP) by 1.33–4.16% (2.6–8.3 mm) and 3.13–8.16% (6.2–18.0 mm) under RCP4.5 and RCP8.5, respectively, relative to the base period (1981–2016). Across the 3H Plain, the mean temperature during the WGP is projected to increase between 1.17–1.21 and 1.17–1.28 °C under RCP4.5 and RCP8.5, respectively. The projections during the maize growth period (MGP) indicate an increase in the temperature between 1.29–1.92 and 1.84–2.08 °C under RCP4.5 and RCP8.5, respectively. During the MGP, precipitation is also projected to increase by 7.41–9.73% (33.6–45.1 mm) and 6.63–14.78% (29.8–72.7 mm) under RCP4.5 and RCP8.5 scenarios, respectively. For each 1% change in climatic factors, the comprehensive effect on yield was projected to be 0.65 and 0.58% for wheat and −1.08 and −1.11% for maize, under RCP4.5 and RCP8.5, respectively, when other factors were kept constant. The change in water resources will be insignificant during the WGP and more pronounced during the MGP. The study provides an overview of changes in meteorological parameters and scientific evidence for climate change adaptation in the wheat–maize cropping system. HIGHLIGHTS The temperature and precipitation are projected to increase under RCP4.5 and RCP8.5 for both winter wheat and summer maize.; The change in climate variables is projected to be beneficial for wheat yield under both emission scenarios with the highest increase in Shandong and parts of Henan.; The summer maize yields are estimated to decrease due to increased temperature, particularly in the north 3H region.;

A good knowledge of future coastal wind and wave resources in the context of climate change is crucial for the construction of offshore wind farms. In this study, the dataset of the coupled model intercomparison project phase 6 (CMIP6) was used to evaluate the future wind resources and wave conditions in the nearshore area of Guangdong of China. The long short-term memory (LSTM) algorithm was used to develop a statistical downscaling method to render high spatial resolution data. The Copula function was used to construct the joint probability distribution function. The key findings were as follows. First, over the whole Guangdong coastal area, the projection of wind speed (Hs) shows a generally increasing trend, the wave height (Ws) remains almost unchanged, and the growth is particularly pronounced in the western area. Second, it is clear that the joint probability distribution is less spread in the late 21st century; therefore, the distribution of wind energy will also be more suitable for the optimal operating of wind turbines. Third, the joint probability distribution results show that large wind and wave conditions possibly occur at the same time, which must be considered when determining worst-case conditions. Future work is required to use more models and scenarios from the ongoing CMIP6. [ABSTRACT FROM AUTHOR]