Statistical Downscaling Model Research Articles

“Climate Change Projections in Jammu and Kashmir under A1B, A2 & B2 Scenarios”

Abstract: The Global Climate Model (GCM) run at a coarse spatial resolution cannot be directly used for climate impact studies. Downscaling is required to extract the sub-grid and local scale information. The present study was undertaken to study the effect of climate change on weather parameters like maximum temperature, minimum temperature and precipitation for A1B, A2 and B2 scenarios. The study uses the Hadley centre coupled model (HadCM3) of the Intergovernmental Panel for Climate Change (IPCC) Forth Assessment Report. Thirty year weather data (1985-2015) obtained form Indian Metrological Department (IMD) Srinagar was used for the study. Statistical Downscaling Model (SDSM), R software and Artificial Neural Network (ANN) models were tested for the Srinagar area. This research investigates which among the three is better model for downscaling climate data for Srinagar area. The modelling results showed a first rate agreement between the experimental data and predicted values for temperature series with high coefficient of determination R2 values varying from (0.93-0.95) for different models. In case of precipitation R2 values varied from (0.08-0.249) for different models. The low values of coefficient of determination in precipitation time series are due to lot of uncertainty. occurrence of precipitation which could not be defined by the selected models. Based on Mean squared error (MSE), Root mean squared error (RMSE), Absolute average deviation (AAD), correlation coefficient and coefficient of determination, the R software performed better than the SDSM and ANN for maximum temperature, minimum temperature & precipitation. Thus R software was used for climate scenario generation. According to our simulated model, precipitation showed a decreasing trend whereas maximum and minimum temperatures showed an increasing trend. An overall increasing pattern of (9.81%) for A1B scenario, (15.24%) for A2 scenario & (10.32%) for B2 scenario for maximum temperature, (2.18%) for A1B scenario, (29.15%) for A2 scenario & (19.90%) for B2 scenario for minimum temperature and an overall decreasing pattern (22.60%) for A1B scenario, (17.23%) for A2 scenario & (7.11%) for B2 scenario for precipitation was noted

Comparison of Generalized Linear Model between Gamma and Tweedie Compound Response for Rainfall Prediction in Lampung Province

Rainfall plays a pivotal role in influencing agricultural production in Lampung province. The precision of rainfall predictions holds significant importance for enhancing agricultural yields in the region. One effective approach for modeling rainfall is Statistical Downscaling (SD), which employs statistical models to examine the correlation between large-scale (global) climatological data and small-scale (local) data. SD addresses the limitation of global scale data, such as the General Circulation Model (GCM), which lacks the resolution to directly forecast localized climate conditions like rainfall. Rainfall can be broadly categorized into continuous and discrete components. The continuous component delineates the intensity of rainfall, while the discrete component describes the occurrence of rain. Both components are integral to accurate rainfall predictions. The mixed Tweedie distribution, combining Gamma and Poisson distributions, is proficient in handling both continuous and discrete components of rainfall. GCMs commonly encounter multicollinearity issues in SD modeling, which can be mitigated through Principal Component Analysis. This study seeks to compare two regression models: the generalized linear model with a gamma response and the Tweedie compound response. Rainfall data from three distinct regions in Lampung province, representing high, medium, and lowlands, is utilized. The research findings indicate that, for high and lowlands, the Tweedie compound exhibits a smaller Root Mean Square Error of Prediction (RMSEP) compared to gamma. Conversely, in medium lands, gamma-GLM demonstrates a smaller RMSEP value than the Tweedie compound. Thus, the distribution of the Tweedie compound is better suited for use than Gamma-GLM, especially for high and lowland areas.

Hydroclimatic projection: statistical learning and downscaling model for rainfall and runoff forecasting

Abstract The study is carried out to investigate the surface runoff depth with changing precipitation due to climate change in the study area where sandy loam and loamy soil are dominant. In this study, future rainfall is projected by a statistical downscaling model (SDSM) using a set of predictors derived from a Coupled Model Intercomparison Project Phase 6 (CMIP6) global climate model (GCM) [the Norwegian Earth System Model (NorESM)] with updated scenarios SSP 4.5 and SSP 8.5. Daily rainfall values for the observed period (1981 to 2014) are validated using statistical learning and evaluated with matrices, namely, root mean square error (RMSE), coefficient of correlation, and Nash–Sutcliffe efficiency (NSE), which are found to be valid for further predictions. Rainfall projections show a decrease in rainfall trend of 50% from 2030 to 2040 for scenario SSP 4.5 and an increase of 7% from 2040 to 2050. Predicted rainfall for scenario SSP 8.5 shows a similar trend of decreasing rainfall of 24% for the period 2030–2040 and an increase of 19% in the period 2040–2050. Furthermore, these rainfall values are spatially modelled in a geographic information system (GIS) and rainfall maps are obtained. The obtained rainfall map, land-use map, and soil map are overlaid to compute curve numbers and runoff depths. A similar trend of decrease in runoff is observed for the period 2030–2050. The overall trend of climate change shows a water-stressed scenario.

A new statistical downscaling approach for short‐term forecasting of summer air temperatures through a fusion of deep learning and spatial interpolation

AbstractReliable early forecasting of extreme summer air temperatures is essential for effectively managing and mitigating the socioeconomic damage caused by thermal disasters. Numerical weather prediction models have become valuable tools for forecasting air temperature; however, they incur high computational costs, resulting in coarse spatial resolution and systematic bias owing to imperfect parametrization. To address these problems, we developed a novel statistical downscaling and bias correction method (named DeU‐Net) for the maximum and minimum air temperature (Tmax and Tmin respectively) forecasts obtained from the Global Data Assimilation and Prediction System with a spatial resolution of 10 to 1.5 km over South Korea through the fusion of deep learning (i.e., U‐Net) and spatial interpolation. In this study, we used a methodology to decompose statistically downscaled Tmax and Tmin forecasts into temporal dynamics over South Korea and spatial fluctuations by pixels. When comparing the proposed DeU‐Net with the dynamical downscaling model (i.e., Local Data Assimilation and Prediction System) and support vector regression‐based statistical downscaling model at the seen and unseen stations for forecasting the next‐day Tmax and Tmin, DeU‐Net showed the highest spatial correlation and the lowest root‐mean‐square error in all cases. In a qualitative evaluation, DeU‐Net successfully produced a detailed spatial distribution most similar to the observations. A further comparison extending the forecast lead time to 7 days indicated that the proposed DeU‐Net is a better downscaling approach than support vector regression, regardless of the forecast lead time. These results demonstrate that bias‐corrected high spatial resolution air temperature forecasts with relatively long forecast lead times in summer can be effectively produced using the proposed model for operational forecasting.

Comparative Study of Observed and Projected Future Climate Evolution in Two Watersheds (Souss-Massa and Ouergha, Morocco) Using the Statistical Downscaling Model (SDSM)

The hydrological cycle and local climate of a region are directly impacted by climate change. Rapid fluctuations in climate alter weather patterns, resulting in the occurrence of extreme weather phenomena. This study examines two basins situated in distinct regions of Morocco, each influenced by specific climatic conditions. The primary aim of this study is to examine the climatic changes occurring within the two watersheds. Initially, it involves a comparative analysis of annual precipitation and temperatures from 1982 to 2022. Subsequently, the study projects precipitation and temperatures for the period spanning from 2024 to 2100. This analysis relies on data collected from 13 stations, 8 in the Souss-Massa and 5 in the Ouergha regions, using the Statistical Downscaling Model (SDSM) version 4.2.9. The investigation employs the fourth generation coupled global climate model CanESM2. Precipitation data and historical temperature models are used to forecast future temperatures and precipitation, based on Representative Concentration Pathway (RCP) scenarios 4.5 and 8.5. The projected scenarios indicate a decrease in observed precipitation for the years 2040, 2060, and 2080, while temperatures are predicted to rise in both watersheds across all future scenarios.

Use of geomatics, Simulating the Impact of Future Land Use and Climate Change on Soil Erosion in the Tigrigra watershed (Azrou region, Middle Atlas, Morocco)

Soil losses need to be quantified in watersheds to implement erosion protection measures. The main objective of this work is to quantify soil loss in the Tigrigra watershed over the reference period 1985-2020 and two future periods 2050-2070, A Revised Universal Soil Loss Equation (RUSLE) model supported by geographic information systems (GIS) and remote sensing was used. GIS’s model generator can automate various operations of creating thematic layers of model parameters. For future climatic periods (2050-2070), precipitation was produced using a classical statistical downscaling model (SDSM). On the other hand, Automata/Markov models (CA Markov) are used to characterize future land use through modeling in Idrisi software. Over the two periods, the results showed that annual erosivity varies decreases, or increases. The annual soil loss maps showed that 50% of our study area was in the very low class (<5 t/ha/year), while 20% was in the severe class (>80 t/ha/year). These fluctuations are primarily due to the effects of climate change and deforestation/reforestation in the region. This leads to changes in soil erosion due to the important role played by these two factors.

Effects of systematic predictor selection for statistical downscaling of rainfall in Hawai'i

AbstractWhile there has been rapid advancement in the development and application of statistical downscaling methods for climate projection, determining the best predictive large‐scale climate information for the targeted local climate variable remains a challenge. The choice of predictor variables is one of the most influential steps of model development and has the potential to lead to varying results, contributing to the total uncertainty of future projections. Despite being a well‐known problem, predictor selection often does not receive adequate attention and the development of a straightforward and feasible prescreening process is needed to provide guidance to simple (e.g., linear regression) and complex (e.g., machine learning) modeling tools alike. In this project, Akaike information criterion (AIC) and leave one out cross validation are used to evaluate sets of predictor variables common in statistical downscaling models of precipitation (e.g., temperature, geopotential height, moisture transport). Because thousands of predictor sets were found to be competitive in their statistical skill in projecting rainfall, results suggest that an ensemble of predictor sets should be used to account for the resulting variance associated with predictor selection. This ensemble approach is applied to make improvements to the most recent statistical downscaling model developed for rainfall projection in Hawai'i. An ensemble validation study is performed by projecting future wet season rainfall in Hawai'i for scenario RCP4.5 using 17 CMIP5 GCMs with ensembles of highly ranked and poorly ranked predictor sets. Results show significantly less variance and improved agreement in the projected sign of precipitation change for the highly ranked predictor set ensemble. In conclusion, it is recommended that statistical downscaling models are run in an ensemble‐mode with multiple combinations of predictors instead of using a single model, allowing for additional quantification of model uncertainty and a “safety net” for when the true physically best‐suited predictor set is unknown.

IWRAM: A hybrid model for irrigation water demand forecasting to quantify the impacts of climate change

The phenomenon of climate change exerts a substantial influence on the water consumption patterns of agricultural crops, thereby affecting the overall demand for irrigation water and regional water security. The quantitative assessment of future changes in irrigation water requirements has great importance for long-term water resource planning and the water-energy-food-ecosystem nexus. This work employed a comprehensive approach (IWRAM) by integrating many models and techniques, including a global climate model (CanESM2), a statistical downscaling model (SDSM), a bias correction technique (QTM), a temperature-based crop phenology method (GDD), and deep learning technology (LSTM), in order to assess the potential impact of climate change on irrigation water demand. Therein, the QTM is utilized for the purpose of bias correction for the CanESM2-SDSMed daily precipitation data. Additionally, the LSTM is implemented to forecast ETo using the CanESM2-SDSMed daily temperature data. The findings from the case study conducted in the Jinghuiqu irrigation area (JIA) indicate that the application of bias correction techniques resulted in notable enhancements in the frequency distribution of predicted precipitation data. Consequently, this led to a more coherent relationship between historical observed data and anticipated future data, aligning them more closely with natural patterns. The deep learning technique exhibited a strong degree of concordance, suggesting its exceptional capacity for modeling daily reference evapotranspiration. By using the IWRAM model, relatively accurate predictions were made for future precipitation, temperature, crop evapotranspiration, and effective precipitation. Consequently, an estimation was made regarding the anticipated need for irrigation water in JIA. These results highlighted that JIA total irrigation water demand (or net irrigation water demand) will increase by 121 mm (net irrigation water: 189 mm) and 119 mm (net irrigation water: 187 mm) in the year 2050 under RCP 2.6 and 8.5 scenarios, respectively. Climate change could potentially affect the seasonal distribution of precipitation. Therefore, it is crucial for the authorities in JIA to not only take into account the overall irrigation water requirement but also give priority to the water requirements of each stage of crop growth.

Climate Change Impacts on Agricultural and Industrial Water Demands in the Beijing–Tianjin–Hebei Region Using Statistical Downscaling Model (SDSM)

As a politically and culturally important city cluster, the Beijing–Tianjin–Hebei (BTH) region is the most prominent area in China where the imbalance between the supply and demand of water resources restricts the sustainable and healthy development of the regional social economy. In the context of global warming, research into water demand prediction that takes climate change into consideration would be more in line with the strategic goal of the low-carbon sustainable development of future cities. At the same time, the prediction of agricultural water demands against a background of climate change is urgently needed, while industrial water consumption is weakly correlated with climate change, an investigation of the statistical relationship between the two is needed. Thus, in this paper, future climate data from the BTH region under the scenarios RCP2.6, RCP4.5 and RCP8.5 were generated using a statistical downscaling model, and then coupled with agricultural and industrial water demand prediction models to simulate and analyze the impact of climate change on the agricultural and industrial water demands, respectively. The results show that during the forecast period (2020–2035), the reference crop evapotranspiration (ET0) growth rates in the Beijing, Tianjin and Hebei areas under the RCP2.6 scenario are 1.438 mm·a−1, 1.393 mm·a−1 and 2.059 mm·a−1, respectively. Under the RCP4.5 scenario, they are 2.252 mm·a−1, 2.310 mm·a−1 and 2.827 mm·a−1, respectively. Under the RCP8.5 scenario, they are 3.123 mm·a−1, 2.310 mm·a−1 and 2.141 mm·a−1, respectively. Furthermore, under each climate scenario, the increase in evapotranspiration in the Hebei area is the largest, followed by that in the Tianjin area, and that in the Beijing area is the smallest. For water consumption per CNY 10,000 of industrial added value during the forecast period, under the three different climate scenarios, a downward trend is seen in the Beijing area, with rates of 0.158, 0.153 and 0.110, respectively, but in the Tianjin area, there is an upward trend, with an upward tendency in rates of 0.170, 0.087 and 0.071, and an upward trend in the Hebei area, with an upward tendency in rates of 0.254, 0.071 and 0.036, respectively. This study will help the BTH region to rationally allocate agricultural and industrial water against the background of future climate change, and strengthen the coordination and cooperation between the different regions to promote the healthy and sustainable development of the cities.

Weather-type statistical downscaling for ocean wave climate in the Chinese marginal seas

A comprehensive understanding of wave characteristics and their variability, based on reliable long-term wave data, is essential in the design, construction, operation and management in offshore and coastal applications. This study aims to downscale multivariate wave data in the Chinese marginal seas at different time scales by employing two weather-type statistical downscaling models. The calibration of both models involves the usage of historical data from ERA5 reanalysis, with sea level pressure and the squared sea level pressure gradient as the predictors and wave parameters as the predictands. The predictor definition considers the swell components in the local wave data. Both models categorize the atmospheric predictors into different weather types using the K-means algorithm (KMA), with and without regression-guided clustering (referred to as WTD-RG and WTD, respectively). Each weather type is associated with sea state parameters. The downscaled wave data from both models are validated against ERA5 data. The results show that both models can generally reproduce the wave statistics for the significant wave height, mean wave direction, mean wave energy flux, and 95th percentile of the significant wave height. However, the WTD-RG model performs better than the WTD model. Additionally, the ERA5 data reveal increasing trends in the mean significant wave height for most areas of the East China Sea and South China Sea and decreasing trends in the Bohai Sea, Yellow Sea, and Beibu Gulf from 1959 to 2021. These patterns are well captured by the WTD-RG model but not by the WTD model.

Impact of climate change on the design of multi-megawatt spar floating wind turbines

Increased frequency and intensity of extreme events can make offshore constructions unsafe due to the rapidly shifting wind-wave pattern. The consequences of climate change are disregarded by the current performance-based design of offshore wind turbines (OWT). The Statistical Downscaling Model (SDSM) and Artificial Neural Network (ANN) algorithm are used to present a simplified approach to enable the inclusion of future climatic projections in the design of spar-floating wind turbines. A two-variable statistical equation employing an Artificial Neural Network is established for calculating wind-induced wave height for the North Sea and West Coast of India, which is a valuable parameter for the site-specific design of offshore constructions. Under the SSP2-4.5 scenario, the North Sea's most likely wind speed is anticipated to decrease by 11 %, whereas the west coast of India experiences a slight decrease in wind speed. Serviceability responses, such as tower deflection, rotation, and nacelle acceleration, are expected to rise by 8–10 %. In contrast, a decrease in these responses is projected in the North Sea due to a decrease in future wind speed and wave height. Climate change has a greater impact on shutdown conditions than on normal operations, primarily due to the pronounced shifts in extreme climate conditions.

Assessing the impact of RCP4.5 and RCP8.5 scenarios on landslide susceptibility mapping using support vector machine: A case study of Penang Island, Malaysia

This paper investigates the climate change influence on landslide susceptibility mapping (LSM) using a case study conducted on Penang Island in Malaysia, a region prone to landslides. This study was carried out due to limited research assessing the climate change effect on LSM, considering rainfall and temperature. The study employs climate factors, including temperature and rainfall, alongside 12 causative factors in a Support Vector Machine (SVM) model to develop LSM. The Statistical Downscaling Model (SDSM) is used to derive climate change projections under two Representative Concentration Pathways (RCP) scenarios (RCP4.5 and RCP8.5). Data preparation and normalization are performed using ArcGIS 10.7. Based on the results, future annual rainfall and daily temperatures are expected to rise under both scenarios, with RCP8.5 exhibiting more significant climatic changes. The LSM zonation is impacted more significantly under RCP8.5 due to the severity of climate change. LSM under an observation period achieves the best results (area under the curve (AUC) = 85.75, average statistical index (SI) = 94.48%, kappa = 0.885), followed by LSM under RCP4.5 (AUC = 84.38, average SI = 93.54%, kappa = 0.865) and LSM under RCP8.5 (AUC = 84.13, average SI = 93.34%, kappa = 0.860), demonstrating their reliability and adequate performance. These LSMs can assist local authorities in designating critical areas for monitoring and implementing an early-warning system to respond more effectively to landslide risks caused by climate change. However, the study's limitation is considering only two climate scenarios (RCP4.5 and RCP8.5). Future research should encompass a broader range of climate scenarios to develop the most reliable LSM, given the high uncertainty associated with climate change.

Projection of Future Climate Change and Its Influence on Surface Runoff of the Upper Yangtze River Basin, China

Global climate change will modify precipitation and temperatures’ temporal and spatial distribution, trigger more extreme weather events, and impact hydrological processes. The Yangtze River basin is one of the world’s largest basins, and understanding future climate changes is vital for water resource management and supply. Research on predicting future climate change in the upper Yangtze River basin (UYRB) and introducing machine learning algorithms to analyze the impact of climate factors, including extreme weather indicators, on surface runoff is urgently needed. In this study, a statistical downscaling model (SDSM) was used to forecast the future climate in the UYRB, and the Mann–Kendall (MK) or modified Mann–Kendall (MMK) trend test at a 5% level of significance was applied to analyze temporal trends. The Spearman rank correlation (SRC) test at a 5% level of significance and random forest regression (RFR) model were employed to identify the key climatic factors affecting surface runoff from annual precipitation, annual temperature, maximum 5-day precipitation (R×5Day), number of tropical nights (TR), and consecutive dry days (CDD), and the RFR model was also used to predict future runoff. Based on the results, we found that, compared to the selected historical period (1985–2014), the mean annual precipitation (temperature) during the mid-term (2036–2065) increased by 18.93% (12.77%), 17.78% (14.68%), 20.03% (17.03%), and 19.67% (19.29%) under SSP1-2.6, SSP2-4.5, SSP3-7.0, and SSP5-8.5, respectively, and during the long term (2071–2100), increased by 19.44% (12.95%), 22.01% (21.37%), 30.31% (30.32%), and 34.48% (37.97%), respectively. The warming and humidification characteristics of the northwestern UYRB were more pronounced. The key climatic factors influencing surface runoff were annual precipitation, maximum 5-day precipitation (R×5day), and annual temperature. Because of warming and humidification, surface runoff in the UYRB is expected to increase relative to the historical period. The surface runoff during the mid-term (long term) increased by 12.09% (12.58%), 8.15% (6.84%), 8.86% (8.87%), and 5.77% (6.21%) under SSP1-2.6, SSP2-4.5, SSP3-7.0, and SSP5-8.5, respectively. The implementation of sustainable development pathways under the low radiative forcing scenario can be effective in mitigating climate change, but at the same time, it may increase the risk of floods in the UYRB.

A Predictive Model For Quantifying The Combined Effect Of Land Use Change And Climatic Variability On Sedimentation In Malaba Sub Catchment

Over the previous years, sedimentation in Malaba sub catchment has become a major problem, this research aimed at using SWAT as a predictive model to assess the impacts of historical and future climate change and land use changes on sedimentation in Malaba sub catchment. Two variables rainfall and temperature were considered under climate change. Four scenarios were modeled in SWAT during this study, the first scenario focused on historical sedimentation in the catchment, land use data of 2000 and weather data of 1990 to 2005 was used to simulate flow and sediment yield in the catchment. The second scenario focused on the current sediment yield in the catchment where land use data of 2020 and weather data of 2003 to 2020 was used to simulate the sediment yield and flow. Calibration and validation was done using SWAT CUP software. The calibrated and validated model was then used to further simulate two future scenarios. The two future scenarios used projected land use data and projected weather data. Land use data was projected using Clerk Labs Terr set software while weather data was projected using the Statistical Downscaling Model (SDSM). The third scenario considered projected land use data of 2035 and projected weather data of 2020 to 2035 to simulate flow and sediment yield of 2023 to 2035. The forth scenario considered projected land use data of 2050 and projected weather data of 2033 to 2050 to simulate catchment flow and sediment yield of 2036 to 2050. The results obtained from climate change analysis indicated un even changes, rainfall over the catchment is expected to increase, maximum temperature is expected to increase, minimum temperature is expected to reduce. Land use change results after comparison between the historical land use and projected land use indicated that from 2020 to 2050 cropland and built up area will increase while forests and water bodies will reduce. Average annual Sediment yield of 1993 to 2020 is expected to generally increase from 4975.4 tons/km2/year to an average annual sediment yield of 5525.6 tons/km2/year from 2023 to 2050. Therefore, mitigation measures should be adopted to ensure sustainable management of the catchment. Some of the recommendations made in this research are; Government through NEMA should put restrictions on cultivation close to the river banks, encourage planting trees within the catchment, bare lands within the catchment should be converted to other sustainable land uses. The study findings of this research shall be relevant for planning, design and management of reservoirs, dams, irrigation systems and sustainability of eco systems in the catchment.

Modeling the influence of daily temperature and precipitation extreme indices on vegetation dynamics in Katsina State using statistical downscaling model (SDM)

Climate change significantly impacts local, regional, and global vegetation changes. These changes have continued to threaten ecosystems, especially in dryland areas where moisture is scarce, and the livelihoods of rural communities are at risk because of such changes, as well as their capacity to provide outputs and sustain the livelihoods of rural communities. This study aims to derive a simulation model of vegetation conditions concerning daily temperature and extreme precipitation indices in Katsina State, Nigeria. This study uses remote sensing and Geographic Information System (GIS)-based analysis, time series analysis of precipitation, maximum and minimum temperature, and twenty-four indices defined by the Expert Team on Climate Change Detection to evaluate the influence of Temp and precipitation extreme indices on vegetation dynamics/NDVI. The results indicate that the statistical downscaling model (SDSM) is performing satisfactorily in predicting maximum and minimum temperatures, along with precipitation, for the time horizon of the simulation. Mean precipitation of 1.98 mm (RCP2.6), 2.03 mm (RCP4.5), and 2.07 mm (RCP8.5) was revealed in the study area. Tmax and Tmin were projected under low emissions at 20.1 0C and 14.25 0C (RCP2.6), 20.13 °C and 14.26 0C (RCP4.5), and 20.15 0C and 14.27 0C (RCP8.5) respectively. All the scenarios present increasing minimum and maximum temperatures and decreasing precipitation, except for RCP8.5, which predicted a more adverse trend. However, in Katsina State, the Normalized Difference Vegetation Index (NDVI), precipitation, Tmax, and Tmin, extreme temperature and precipitation indices, and drought indices successfully demonstrated spatial and temporal variability.

Leveraging Deep Learning for Climate Change Prediction Models : A Dive into Cutting-Edge Techniques

In recent years, deep learning has emerged as a promising approach for statistical downscaling, which involves generating high-resolution climate data from coarse low-resolution variables. However, concerns persist regarding the ability of these models to generalize to future climate change scenarios, primarily due to the assumption of stationarity. In this study, we propose the use of deep ensembles as a straightforward method to enhance the uncertainty quantification of statistical downscaling models. By improving the representation of uncertainty, these models offer superior planning capabilities against extreme weather events, which can have significant negative social and economic impacts. Given the absence of observational future data, we rely on pseudo-reality experiments to evaluate the effectiveness of deep ensembles in quantifying the uncertainty of climate change projections. The adoption of deep ensembles facilitates more robust risk assessment, addressing critical needs in various sectors for adapting to climate change challenges.

PyClim-SDM: Service for generation of statistically downscaled climate change projections supporting national adaptation strategies

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 intervariableinconsistencies 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.

Evaluation of statistical downscaling model's performance in projecting future climate change scenarios

Abstract 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.

A high-resolution daily global dataset of statistically downscaled CMIP6 models for climate impact analyses

A large number of historical simulations and future climate projections are available from Global Climate Models, but these are typically of coarse resolution, which limits their effectiveness for assessing local scale changes in climate and attendant impacts. Here, we use a novel statistical downscaling model capable of replicating extreme events, the Bias Correction Constructed Analogues with Quantile mapping reordering (BCCAQ), to downscale daily precipitation, air-temperature, maximum and minimum temperature, wind speed, air pressure, and relative humidity from 18 GCMs from the Coupled Model Intercomparison Project Phase 6 (CMIP6). BCCAQ is calibrated using high-resolution reference datasets and showed a good performance in removing bias from GCMs and reproducing extreme events. The globally downscaled data are available at the Centre for Environmental Data Analysis (https://doi.org/10.5285/c107618f1db34801bb88a1e927b82317) for the historical (1981–2014) and future (2015–2100) periods at 0.25° resolution and at daily time step across three Shared Socioeconomic Pathways (SSP2-4.5, SSP5-3.4-OS and SSP5-8.5). This new climate dataset will be useful for assessing future changes and variability in climate and for driving high-resolution impact assessment models.

Comparison between statistical and dynamical downscaling of rainfall over the Gwadar‐Ormara basin, Pakistan

AbstractThis 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.