Statistical Downscaling Model Research Articles

Assessing vulnerability to climate change for total organic carbon in a system of drinking water supply

Evaluation of the impact of climate change on water sectors is one of the most important challenges of water resources management in the current century. In this paper, assessment of vulnerability to climate change on total organic carbon (TOC) concentration in the Jajrood catchment in Tehran has been done. The hydroclimatologic parts of the assessment have been done by implementing HYMOD and downscaling of precipitation, temperature and evaporation using statistical downscaling model (SDSM). Three climate change scenarios (RCP 2.6, RCP 4.5 and RCP 8.5) from the Canadian Earth System Model have been used to project climate change (2006 to 2050). Moreover, different linear models were used to model water quality in Latyan Dam and Tehranpars Water Treatment Plant (TWTP). Based on the results, the average annual simulated rainfall under RCP 2.6, RCP 4.5, and RCP 8.5 are 670.2, 658.06 and 650.93 mm, respectively. The projected streamflow under the scenarios were 97%, 90%, and 92%, respectively, more than observation period. The average simulated TOC concentration in influent water was 52.43 (mg/L), which was about seven times more than the averaged observed value (8.78 mg/L). The most important byproduct of the current impact assessment is to define an adaptation action plan, because the TWTP will not be able to provide drinking water standards in this area.

Comparative study on the reservoir operation planning with the climate change adaptation

The management planning of Pedu–Muda reservoir, Kedah, was investigated in the context of the climate change evolution. The aim of this study was to evaluate the impact of the climate change to the reservoir operating management system and its sustainability. The study was divided into two sections; Analysis 1 refers to the reservoir optimization adapted with the climate assessment. The statistical downscaling model reacted as the climate model to generate the long-term pattern of the local climates affected by the greenhouse gases. Analysis 2 refers to the reservoir optimization but excluded the climate changes assessment in the analyses. The non-dominated sorting genetic algorithm version II (NSGA-II) was applied in both analyses to optimize the water use due to the multi-objectives demand, maximizing water release, minimizing water shortage and maximizing reservoir storage. The formation of Pareto optimal solutions from both analyses was measured and compared. The results showed the Analysis 1 potential to produce consistence monthly flow with lesser error and higher correlation values. It also produced better Pareto optimal solution set and considered all the objectives demands. The NSGA-II also successfully improves and re-manages the reservoir storage efficiently and reduce the dependancy of these reservoirs.

Downscaling Precipitation in the Data-Scarce Inland River Basin of Northwest China Based on Earth System Data Products

Precipitation is a key climatic variable that connects the processes of atmosphere and land surface, and it plays a leading role in the water cycle. However, the vast area of Northwest China, its complex geographical environment, and its scarce observation data make it difficult to deeply understand the temporal and spatial variation of precipitation. This paper establishes a statistical downscaling model to downscale the monthly precipitation in the inland river basin of Northwest China with the Tarim River Basin (TRB) as a typical representation. This method combines polynomial regression and machine learning, and it uses the batch gradient descent (BGD) algorithm to train the regression model. We downscale the monthly precipitation and obtain a dataset from January 2001 to December 2017 with a spatial resolution of 1 km × 1 km. The results show that the downscaling model presents a good performance in precipitation simulation with a high resolution, and it is more effective than ordinary polynomial regression. We also investigate the temporal and spatial variations of precipitation in the TRB based on the downscaling dataset. Analyses illustrate that the annual precipitation in the southern foothills of the Tianshan Mountains and the North Kunlun Mountains showed a significant upward trend during the study periods, while the annual precipitation in the central plains presented a significant downward trend.

The potential of canonical correlation analysis in multivariable screening of climate model

The statistical downscaling model (SDSM) been used to analyse the potential changes of local climate trend in the long term. The difficulty of the SDSM model in selecting the best predictors group which having good association to the local climate. Even the SDSM provides screening process to analyse the predictor-rainfall relationship, however it has limited ability in analysing multiple variables from 26 predictors with 10 rainfall stations around Kedah state, Malaysia. In this regard, the Canonical Correlation Analysis (CCA) been used to analyse the multi predictor-rainfall relationships. The concept of canonical coefficient is sufficient to show the capability and reliability of the predictors based on the percentages of variance that can explained in the dependent variable using the independent variable. There were 10 predictors’ group have been developed and one predictor’s group was built based on the CCA result. The performances of these predictors groups were tested using statistical analyses. Results revealed that the predictors group selected by the CCA method has produced smaller values of MAE and MSE for all stations except at station of Ladang Tanjung Pauh. The box plot’s results, which generated from one hundred simulated samples, indicated that the performance of CCA method was remarkable. The presence of discrepancies in the HadCM3-A2 and HadCM3-B2 scenario simulations were relatively small and considered acceptable.

Liu-type regression in statistical downscaling models for forecasting monthly rainfall salt as producer regions in Pangkep regency

Liu-Type Regression (LTR) is one of the statistical methods to overcome multicollinearity in multiple regression models. LTR is the development of Ridge regression and Liu estimator. When there is a strong collinearity, selected k parameter in the ridge regression does not fully overcome the multicollinearity. This study aimed to estimate the rainfall data in Pangkep Regency (as response variable) with LTR approach on Statistical Downscaling (SD) models. Precipitation (as predictor variables) is the result of a simulation of a grid on the Global Circulation Model (GCM). This study uses a size 8 8 grid of GCM (64 predictor variables) over an area of Pangkep Regency so that there is a high multicollinearity. Three dummy variables were determined from k-means cluster technique used as predictor variables to overcome the heterogeneity of residual variance. LTR model with dummy variables are able to explain the diversity of rainfall data properly. The value of R2 produced ranges 85.23% -88.99% with Root Mean Square Error (RMSE) ranges 117.732-136.377. Validation of the model generates a high correlation value between the actual rainfall and alleged rainfall period of 2017 (about 0.977-0.979). The value of Root Mean Square Error Prediction (RMSEP) produced lower (about 57.625-61.120). SD analysis was also performed with and without the dummy variable in the Ridge regression and LTR. In general, LRT models with dummy (k = 0.652, d = -0.799) is the best model based on the value of R2, RMSE, correlation, and RMSEP.

Downscaling and Projection of Spatiotemporal Changes in Temperature of Bangladesh

The objectives of this study were to: (1) evaluate possible deviations in annual and seasonal maximum (Tmx) and minimum (Tmn) temperatures, and, (2) determine the spatial pattern of these temperature changes. The study used statistical downscaling of the Coupled Model Intercomparison Project Phase 5 (CMIP5) global climate models (GCMs) under four representative concentration pathway (RCP) scenarios. Perfect prognosis statistical downscaling models, based on support vector machine (SVM), were developed for this purpose. Biases in the GCM simulations were corrected using quantile mapping (QM) and the data were then used to determine future temperature scenarios at different locations within Bangladesh. For most of the GCMs, the mean bias was close to zero and the Nash–Sutcliffe efficiency was above 0.58 for the downscaled temperature. Non-parametric hypothesis tests showed equality in median, distribution and variance values of the observed and downscaled temperature for all GCMs. Temperature projections from the models revealed an increase in Tmx by 1.3–2.3 °C, 1.3–2.9 °C, 1.5–3.1 °C, and 2.2–4.3 °C, and Tmn by 1.8–3.0 °C, 2.1–4.2 °C, 2.4–4.5 °C, and 3.2–5.1 °C under the four RCPs during the 2070–2099 period when compared with the 1971–2000 period. The greatest increase in Tmx and Tmn was found in the more northern regions and the lowest increase was found in the southeast coastal region. Tmn tended to increase in winter, while Tmx increased predominantly during summer. Uncertainty in the temperature projections was found to be greater during the latter part of the century. The rapid rise in temperature predicted for the northern part of Bangladesh (which is historically prone to temperature extremes) may cause an increase in the frequency of temperature-related extremes in this region.

Assessment of climate change impact on surface runoff, statistical downscaling and hydrological modeling

Over the last few decades, the climate change has been increased due to the increased industrial activities, greenhouse gas emissions, and CO2 level. This change has affected the water resource management so that the amount of water entered from upstream of watersheds has been transformed every year, and the water resource management has become difficult for the surface runoff, entered water, flood and drought. The problem becomes more serious when the study area (Kan watershed) is located upstream of such urban watershed as Tehran, where the climate change studies on the water resources are very important. In this study, using the Statistical downscaling model (SDSM), the data of CanESM2 Canadian general circulation model (GCM) was downscaled under the Representative Concentration Pathway (RCP) 2.6, RCP4.5, and RCP8.5. In order to study the climate change, the artificial neural network (ANN) and IHACRES models were used over the period of 2010–2040. The study results showed that the temperature is increased in the upcoming period of 2006–2100 (0.8–5.6 C°), and the highest temperature changes are related to winter and summer. The precipitation in the upcoming period shows an increasing trend on the annual average, but, in general, it can be said that the 4–55% precipitation shows an increasing trend. The runoff in the upcoming period of 2010–2040 under RCP2.6, RCP4.5 and RCP8.5 is −4, 26 and −2 percent in the ANN model and 26, 28 and 33 percent in the IHACRES model, respectively.

Incorporation of Long Term Climate Changes in Hydrological Modelling

One of climate change's most important concerns at the moment is its impact on hydrology as it has direct links with agriculture, vegetation, and livelihood. This study tries to analyze potential future climate change in the Kumaradhara river basin. This study involved three steps: (1) acquiring and using general circulation model (GCM) to project future global climate scenarios; (2) establishing statistical relationships between GCM data and observed data using Statistical Downscaling Model (SDSM); (3) downscaling the second generation Canadian Earth system Model (CanESM2)GCM output based on the established statistical relationship. The statistical downscaling is carried out for three scenarios used in the fifth evaluation report of the recent Intergovernmental Panel on Climate Change (IPCC) viz., Representative Concentration Pathways (RCPs) 2.6, 4.5 and 8.5. The statistical downscaling Model (SDSM) results showed that the mean annual daily precipitation is altered in the basin under all the scenarios but it will be different in different time periods depending on scenarios and the basin will experience the reduced precipitation levels in summer. Also the precipitation will marginally rise in all the time slices with reference to baseline data. We can conclude from the results that this region's climate will affect future farming as the availability of water is bound to change. This study should, however, be followed up by a larger study incorporating multiple CMIP5 models such that changes in hydrological-regimes can be examined appropriately.

A novel approach for modelling pattern and spatial dependence structures between climate variables by combining mixture models with copula models

AbstractSpatiotemporal dependence structures play a pivotal role in understanding the meteorological characteristics of a basin or subbasin. This further affects the hydrological conditions and, consequently, will provide misleading results if these structures are not taken into account properly. In this study, we modelled the spatial dependence structure of three climate variables, maximum and minimum temperature and precipitation, throughout the Monsoon‐dominated zone of Pakistan. For temperature, six meteorological stations have been considered, for precipitation we used the results of four meteorological stations. For modelling the dependence structure between temperature and precipitation at multiple sites, we utilized C‐Vine, D‐Vine and student t‐copula models. For temperature, multivariate mixture normal distributions, and for precipitation, the gamma distribution, have been used as marginals under the copula models. The models were calibrated by utilizing the 20 years daily data from 1981 to 2000, and for validation, we used the data for 10‐year period from 2001 to 2010. The simulations were performed for each variable separately, conditioned on spatial neighbours. A comparison was made between the different copula models, on the basis of observational and simulated patterns and spatial dependence structures, the performance was evaluated for the validation period. The results show that all copula models performed well; however, there are subtle differences between them. The copula models captured the patterns and spatial dependence structures between climate variables, however, the t‐copula showed poor performance in reproducing the dependence structure with respect to magnitude. It was observed that important statistics of observed data have been closely approximated except a few maximum values for maximum temperature and minimum values for minimum temperature. Probability density functions of simulated data follow closely the pattern of observational data. These methods can be combined with statistical downscaling models to improve their performance, particularly in modelling the dependence structure between climate variables at multiple sites.

Long-Term Perspective Changes in Crop Irrigation Requirement Caused by Climate and Agriculture Land Use Changes in Rechna Doab, Pakistan

Climate change and agriculture land use changes in the form of cropping patterns are closely linked with crop water use. In this study the SDSM (statistical downscaling model) was used to downscale and simulate changes in meteorological parameters from 1961 to 2099 using HadCM3 General Circulation Model (GCM) data under two selected scenarios i.e., H3A2 and H3B2. Results indicated that Tmax, Tmin, and wind speed may increase while relative humidity and precipitation may decrease in the future under both H3A2 and H3B2 scenarios. Downscaled meteorological parameters were used as input in the CROPWAT model to simulate crop irrigation requirement (CIR) in the baseline (1961–1990) and the future (2020s, 2050s and 2080s). Data related to agriculture crop sown area of five major crops were collected from Punjab statistical reports for the period of 1981–2015 and forecasted using linear exponential smoothing based on the historical rate. Results indicated that the cropping patterns in the study area will vary with time and proportion of area of which sugarcane, wheat, and rice, may exhibit increasing trend, while decreasing trend with respect to the baseline scenario was found in maize and cotton. Crop sown area is then multiplied with CIR of individual crops derived from CROPWAT to simulate Net-CIR (m3) in three sub-scenarios S1, S2, and S3. Under the H3A2 scenario, total CIR in S1, S2, and S3 may increase by 3.26 BCM, 12.13 BCM, and 17.20 BCM in the 2080s compared to the baseline, while under the H3B2 scenario, Net-CIR in S1, S2, and S3 may increase by 2.98 BCM, 12.04 BCM, and 16.62 BCM in the 2080s with respect to the baseline. It was observed that under the S2 sub-scenario (with changing agriculture land-use), total CIR may increase by 12.13 BCM (H3A2) and 12.04 BCM (H3B2) in the 2080s with respect to the baseline (1961–1990) which is greater as compared to S1 (with changing climate). This study might be valuable in describing the negative effects of climate and agriculture land use changes on annual crop water supply in Rechna Doab.

An evaluation of two statistical downscaling models for downscaling monthly precipitation in the Heihe River basin of China

The stepwise regression model (SRM) is a widely used statistical downscaling method that could be used to establish a statistical relationship between observed precipitation and predictors. However, the SRM cannot reflect the contributions of predictors to precipitation reasonably, which may not be the best model based on several possible competing predictors. Bayesian model averaging (BMA) is a standard inferencing approach that considers multiple competing statistical models. The BMA infers precipitation predictions by weighing individual predictors based on their probabilistic likelihood measures over the training period, with the better-performing predictions receiving higher weights than the worse-performing ones. Furthermore, the BMA provides a more reliable description of all the predictors than the SRM, leading to a sharper and better calibrated probability density function (PDF) for the probabilistic predictions. In this study, monthly precipitation at fifteen meteorological stations over the period of 1971–2012 in the Heihe River basin (HRB), which is located in an arid area of Northwest China, was simulated using the Bayesian model averaging (BMA) and the stepwise regression model (SRM), which was then compared with the observed datasets (OBS). The results showed that the BMA produced more accurate results than the SRM when used to statistically downscale large-scale variables. The multiyear mean precipitation results for twelve of the fifteen meteorological stations that were simulated by the BMA were better than those simulated by the SRM. The RMSE and MAE of the BMA for each station were lower than those of the SRM. The BMA had a lower mean RMSE (− 13.93%) and mean MAE (− 14.37%) compared with the SRM. The BMA could reduce the RMSEs and MAEs of precipitation and improve the correlation coefficient effectively. This indicates that the monthly precipitation simulated by the BMA has better consistency with the observed values.

Characterising uncertainty in precipitation downscaling using a Bayesian approach

Statistical downscaling of GCM simulations is widely used for examining future changes in precipitation at different spatial and temporal scales. However, the downscaling process is affected by uncertainty associated with the downscaling model, its parameters, and also the use of different reanalysis products for model calibration. This study develops a Bayesian approach to calibrating a statistical downscaling model. The study investigates the impact of using two different reanalysis products, the National Centre for Environmental Prediction/National Centre for Atmospheric Research Reanalysis 2 (NCEP2) and the European Centre for Medium-Range Forecasts Interim Reanalysis (ERAI), in precipitation downscaling over the Tibetan plateau, a region with sparse ground precipitation coverage. The selected reanalysis products are used for modelling precipitation at selected locations with long, high quality records and diverse geographic characteristics. An assessment of the downscaled precipitation results using atmospheric variables from the ACCESS 1.3 GCM to drive the downscaling model calibrated using a reanalysis dataset is also performed and the impact of calibration uncertainty quantified. The outcomes of this study reveal that the choice of the data length used and the type of Reanalysis product adopted have a significant effect in downscaled precipitation characteristics and their uncertainties, such as the wetness fraction and average annual precipitations over the study locations. These findings point to a common problem in statistical downscaling applications and one that has not been recognised until now. The results show that downscaling model considering ERAI reproduce observed precipitation attributes to a better extent than NCEP2.

Susceptibility of Hydropower Generation to Climate Change: Karun III Dam Case Study

Climate change can cause serious problems for future hydropower plant projects and make them less economically justified. Changing precipitation patterns, rising temperatures, and abrupt snow melting affect river stream patterns and hydropower generation. Thus, study of climate change impacts during the useful life of a hydropower dam is essential and its outcome should be considered in assessing long-term dam feasibility. The aim of this research is to evaluate the impacts of climate change on future hydropower generation in the Karun-III dam located in the southwest region of Iran in two future tri-decadal periods: near (2020–2049) and far (2070–2099). Had-CM3 general circulation model predictions under A2 and B2 SRES scenarios were applied, and downscaled by a statistical downscaling model (SDSM). An artificial neural network (ANN) and HEC-ResSim reservoir model respectively simulated the rainfall–runoff process and hydropower generation. The projections showed that the Karun-III dam catchment under the two scenarios will generally become warmer and wetter with a slightly larger increase in annual precipitation in the near than the far future. Runoff followed the precipitation trend by increasing in both periods. The runoff peak also switched from April to March in both scenarios, due to higher winter precipitation, and earlier snowmelt, which was caused by temperature rise. According to both scenarios, hydropower generation increased more in the near future than in the far future. Annual average power generation increased gradually by 26.7–40.5% under A2 and by 17.4–29.3% under B2 in 2020–2049. In the far period, average power generation increased by 1.8–8.7% in A2 and by 10.5–22% under B2. In the near future, A2 showed energy deduction in the months of June and July, while B2 revealed a decrease in the months of April and June. Additionally, projections in the 2070–2099 under A2 exhibited energy reduction in the months of March through July, while B2 revealed a decrease in April through July. The framework utilized in this study can be exploited to analyze the susceptibility of hydropower production in the long term.

Impacts of projected change in climate on water balance in basins of East Africa

In East Africa, climate change and variability have shown a strong impact on sectors such as agriculture, energy, and water. To allow mitigation and adaptation of the possible impacts of the projected change in climate, this study applies a Statistical Downscaling Model (SDSM) to generate a high-resolution climate projection, equivalent to future station data, to drive impact assessment models in selected, agricultural intensive, basins of Ethiopia (EthShed), Kenya (KenShed), and Tanzania (TanShed). Observed and large-scale climate variables (predictors) are obtained from the national meteorological agency of Ethiopia and international databases. BROOK90, a physical-based hydrological model, is used to assess the impacts of the projected change in precipitation and maximum and minimum temperature (T-max, and T-min) on the water balance. Based on SDSM, the results show an increase in precipitation, relative to the baseline period (1961–1990), in EthShed (14% - 50%) and KenShed (15% - 86%) and a decrease in TanShed (1.3% - 6.3%) in the 20s (2011–2040), 50s (2041–2070), and 80s (2071–2100) under the three Representative Concentration Pathways (RCPs; RCP2.6, RCP4.5, and RCP8.5). T-max (anomalies up to 3.7 °C) and T-min (anomalies up to 2.76 °C) will be warmer than the baseline period throughout the 21 century in all three basins. In line with the projected change in precipitation and temperature, an increase (decrease) in seasonal and annual streamflow, soil-water, and evaporation in EthShed and KenShed (TanShed) is projected in the 20s, 50s, and 80s. In general, sustainable adaptation measures are required to be developed in a site-specific manner, considering the projected increase in temperature and evaporation in all three basins and a decrease in soil-water and streamflow in TanShed.

Assessing the impact of climate change on rainwater harvesting in the Oum Zessar watershed in Southeastern Tunisia

Climate change is believed to have a large impact on water resources system both globally and regionally. It has become a major global issue, especially in developing countries because these are most affected by its impacts. Rainwater harvesting techniques offer an alternative source of water and represent specific adaptive strategies to cope with water scarcity within future climate change. Studying the impact of climate change on rainwater harvesting techniques, however, is difficult, because the general circulation models (GCMs) which are widely used to simulate scenarios of future climate change operate on a coarse scale. We estimated the impact of climate change on water availability at the watershed level by downscaling precipitation and temperature from the GCMs using a statistical downscaling model. A water harvesting model then assessed the performance of the rainwater harvesting techniques for the Oum Zessar watershed in southeastern Tunisia under current climatic conditions and scenarios of future climate change. Annual temperature tended to increase and precipitation tended to decrease. These changes of climatic variables were used in the water harvesting model to simulate future water availability. Changing the directions of water flow between sub-catchments in combination with changing the spillway heights strongly affected the performance of rainwater harvesting under the scenarios of future climate, resulting in a sufficient water supply for 92% of all sub-catchments, compared to 72% without these changes.

Comparison of dynamical and statistical rainfall downscaling of CMIP5 ensembles at a small urban catchment scale

Downscaling of future projections of climatic variables from global climate models (GCMs) to urban catchment scales is required for stormwater studies as existing GCMs are unable to predict the rainfall at high temporal and spatial resolutions. In the current study, the capability of dynamical and statistical downscaling methods is evaluated and compared using the rainfall data of current climate and future extreme events at a small urban catchment scale. Regional climate model (RCM) and statistical downscaling model (SDSM) are utilized to downscale rainfall data from twelve GCMs under three representative concentration pathways (RCPs) (i.e., RCP2.6, RCP 4.5 and RCP 8.5). The daily rainfall data (1985–2015) of Lucas Creek catchment located in Auckland, New Zealand is used as a baseline/current climate. The future downscaled rainfall data is analyzed in the 2090s (2071–2100). The results showed that both the methods performed well in downscaling the current climate. For future projections, SDSM underestimated mean daily rainfall at the start of the annual cycle and overestimated towards the middle of the year compared to RCM. Similarly, monthly variance and skewness were overestimated for some months by SDSM. The GCMs of both the methods also showed variations in the future rainfall projections amongst themselves. However, significant alterations in the future rainfall were observed compared to the current climate. Rainfall frequency analysis was performed by applying the Gumbel distribution to the baseline and downscaled data for 1, 2, 3, 4, 5, 10, 20, 30, 50 and 100 years return periods. The investigation revealed that RCM and SDSM show similar results for low return periods and different results for high return periods for the current and future climate. Both the methods forecasted an increase in magnitudes of future events however, RCM projections were lower compared to SDSM. The results illustrate the downscaling abilities of both the methods at a small urban catchment scale however, the contrasting implication in downscaling the rainfall data is related to their different downscaling mechanisms.

The influence of dependence in characterizing multi-variable uncertainty for climate change impact assessments

ABSTRACTFew approaches exist that explicitly use the uncertainty associated with the spread of climate model simulations in assessing climate change impacts. An approach that does so is second-order approximation (SOA). This incorporates quantification of uncertainty to ascertain its impact on the derived response using a Taylor series expansion of the model. This study uses SOA in a statistical downscaling model of monthly streamflow, with a focus on the influence of dependence in the uncertainty of multiple atmospheric variables. Uncertainty is quantified using the square root error variance concept with a new extension that allows the inter-dependence terms among the atmospheric variable uncertainty to be specified. Applying the model to selected point locations in Australia, it is noted that the downscaling results differ considerably from downscaling that ignores uncertainty. However, when the effects of dependence in uncertainty are incorporated, the results differ according to the regional variations in dependence structure.

Statistically downscaled climate dataset for East Africa

For many regions of the world, current climate change projections are only available at coarser spatial resolution from Global Climate Models (GCMs) that cannot directly be used in impact assessment and adaptation studies at regional and local scale. Impact assessment studies require high-resolution climate data to drive impact assessment models. To overcome this data challenge, we produced a station based climate projection (precipitation and maximum and minimum temperature) for Ethiopia, Kenya, and Tanzania using observed daily data from 211 stations obtained from the National Meteorological Agency of Ethiopia and international databases. Moreover, 26 large-scale climate variables derived from the National Centers for Environmental Prediction reanalysis data (1961–2005) and second generation Canadian Earth System Model (CanESM2, 1961–2100) are used. Statistical Down-Scaling Model (SDSM) is used to produce the required high-resolution climate projection by developing a statistical relationship between the large- and local-scale climate variables. The predictors are analysed more than 16458 times and we provided 20 ensembles for the current (1961–2005) and future (2006–2100, under RCP2.6, RCP4.5, and RCP8.5) climate.

Euro-Cordex Regional Projection Models: What Kind of Agreement for Europe?

EURO-CORDEX is an international initiative which provides regional climate projections based on multiple dynamical and empirical–statistical downscaling models. The goal of this work is to analyse the agreement between projections of the CLMCOM-CCLM4-8-17 (CLMCOM) and SMHI-RCA4 (SMHI) models across Europe. The variables temperature, precipitation and solar radiation were considered for a historical period (1971–2005) and for a future scenario (2006–2040). The overall agreement ($$\mathcal {A}$$) is defined as the normalised area of the magnitude-squared coherence function over the frequency range (averaged over time), being 0 for no agreement and 1 for total agreement between models. The relative mean difference ($$\mathcal {M}$$) and difference between the coefficients of variation ($$\mathcal {V}$$) are also explored, since coherence analysis cannot evaluate differences in mean and variability. Agreement values ranging from 0.32 to 0.74, 0.28 to 0.69 and 0.32 to 0.58 were found for temperature, precipitation and solar radiation, respectively, for the historical period. In all cases, the results show better agreement for lower than higher frequencies. Overall, the time series from both models behave fairly similarly for lower frequencies (i.e. the trend of the time series), while for higher frequencies (i.e. rapid changes in the time series), the similarities between the models are less consistent. For temperature, the $$\mathcal {M}$$ and $$\mathcal {V}$$ values are smaller than 2.5%, while for precipitation and solar radiation they can exceed 50% and 35%, respectively. Further analysis revealed that the contribution of winter and summer differs considerably for $$\mathcal {M}$$ and $$\mathcal {V}$$ values. In conclusion, it seems that such models can provide fairly similar results when considering long periods of time.

Pros and cons of using wavelets in conjunction with genetic programming and generalised linear models in statistical downscaling of precipitation

Among the regression techniques used in building statistical downscaling models, genetic programming (GP) which mimics Darwin’s theory of biological evolution possesses several pros such as it evolves explicit linear or non-linear relationships while identifying optimum predictors, and it discards irrelevant and redundant information in predictors. However, GP is known to simulate unphysically large outliers of predictands. In statistical downscaling, decomposition of predictand and predictor data into number of different time-frequency components with wavelet transform and modelling each component separately should better simulate the time-frequency properties of the predictand, in theory. Therefore, it is important to investigate pros and cons of using GP with wavelet transform in building downscaling models. In this study, wavelet and non-wavelet-based precipitation downscaling models were developed employing GP and generalised liner models (GLM) for 50 stations located in wet and dry climate, with 20CR and NCEP/NCAR reanalysis data, for the investigation of the above matter. It was found that regardless of the mother wavelet, vanishing moment and climate regime, with the increase in decomposition level, the wavelet-based downscaling models developed with GLM tended to show a distinct deterioration in performance in both calibration and validation unlike the wavelet-based downscaling models developed with GP. This was because GP is able to discard unnecessary/redundant information flowing into the model with the increase in the decomposition level through evolution. Furthermore, it was found that when GP is coupled with wavelet transform, the simulation of unphysically large values of the predictand increases significantly.