Maximum Temperature Forecasts Research Articles

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.

Comparison between the Operational and Statistical Daily Maximum and Minimum Temperature Forecasts on the Central Coast of Peru

Abstract Multiple linear regression models were developed for 1–3-day lead forecasts of maximum and minimum temperature for two locations in the city of Lima, on the central coast of Peru (12°S), and contrasted with the operational forecasts issued by the National Meteorological and Hydrological Service—SENAMHI and the output of a regional numerical atmospheric model. We developed empirical models, fitted to data from the 2000–13 period, and verified their skill for the 2014–19 period. Since El Niño produces a strong low-frequency signal, the models focus on the high-frequency weather and subseasonal variability (60-day cutoff). The empirical models outperformed the operational forecasts and the numerical model. For instance, the high-frequency annual correlation coefficient and root-mean-square error (RMSE) for the 1-day lead forecasts were 0.37°–0.53° and 0.74°–1.76°C for the empirical model, respectively, but from around −0.05° to 0.24° and 0.88°–4.21°C in the operational case. Only three predictors were considered for the models, including persistence and large-scale atmospheric indices. Contrary to our belief, the model skill was lowest for the austral winter (June–August), when the extratropical influence is largest, suggesting an enhanced role of local effects. Including local specific humidity as a predictor for minimum temperature at the inland location substantially increased the skill and reduced its seasonality. There were cases in which both the operational and empirical forecast had similar strong errors and we suggest mesoscale circulations, such as the low-level cyclonic vortex over the ocean, as the culprit. Incorporating such information could be valuable for improving the forecasts. Significance Statement We wanted to compare the temperature of the operational forecast of the Meteorological and Hydrological Service, an atmospheric model, and persistence with the observed temperatures on the Peruvian central coast. In addition, we generated an empirical forecast model considering both atmospheric and local predictors. We got better results with this empirical model, considering the highest Pearson correlations and the lowest RMSE values. These results will allow us to use this empirical model as the main tool to automate the forecast on the central coast of Peru. Future work should be aimed at testing the skill of this model for forecasting in other cities of Peru.

Improvement of Maximum Air Temperature Forecasts Using a Stacking Ensemble Technique

Due to the influence of complex factors such as atmospheric dynamic processes, physical processes and local topography and geomorphology, the prediction of near-surface meteorological elements in the numerical weather model often has deviation. The deep learning neural networks are more flexible but with high variance. Here, we proposed a stacking ensemble model named FLT, which consists of a fully connected neural network with embedded layers (ED-FCNN), a long short-term memory (LSTM) network and a temporal convolutional network (TCN) to overcome the high variance of a single neural network and to improve prediction of maximum air temperature. The case study of daily maximum temperature forecast evaluated with observation of almost 2400 weather stations shows substantial improvement over that of single neural network model, ECMWF-IFS and statistical post-processing model. The FLT model can more effectively improve the forecast bias of the ECMWF-IFS model than that of any of the above single neural network model, with the RMSE reduced by 52.36% and the accuracy of temperature forecast increased by 43.12% compared with the ECMWF-IFS model. The average RMSEs of the FLT model decreases by 8.39%, 1.50%, 2.96% and 16.03%, respectively, compared with ED-FCNN, LSTM, TCN and the decaying average method.

Spatial correlation in weather forecast accuracy: a functional time series approach

A functional time series approach is proposed for investigating spatial correlation in daily maximum temperature forecast errors for 111 cities spread across the U.S. The modelling of spatial correlation is most fruitful for longer forecast horizons, and becomes less relevant as the forecast horizon shrinks towards zero. For 6-day-ahead forecasts, the functional approach uncovers interpretable regional spatial effects, and captures the higher variance observed in inland cities versus coastal cities, as well as the higher variance observed in mountain and midwest states. The functional approach also naturally handles missing data through modelling a continuum, and can be implemented efficiently by exploiting the sparsity induced by a B-spline basis. The temporal dependence in the data is modeled through temporal dependence in functional basis coefficients. Independent first order autoregressions with generalized autoregressive conditional heteroskedasticity [AR(1)+GARCH(1,1)] and Student-t innovations work well to capture the persistence of basis coefficients over time and the seasonal heteroskedasticity reflecting higher variance in winter. Through exploiting autocorrelation in the basis coefficients, the functional time series approach also yields a method for improving weather forecasts and uncertainty quantification. The resulting method corrects for bias in the weather forecasts, while reducing the error variance.

Forecasting Maximum Mechanism Temperature in Advanced Technology Microwave Sounder (ATMS) Data Using a Long Short-Term Memory (LSTM) Neural Network

Among the monitored telemetry raw data record (RDR) parameters with the STAR Integrated/Validation System (ICVS), the Advanced Technology Microwave Sounder (ATMS) scan motor mechanism temperature is especially important because the instrument might be unavoidably damaged if the mechanism temperature exceeds 50 °C. In the current operational flight processing software, the instrument automatically enters safe mode and stops collecting scientific data whenever the mechanism temperature exceeds 40 °C. This approach inevitably leads to the instrument entering safe mode unnecessarily at a premature time, causing the loss of scientific data before the mechanism temperature reaches 50 °C. This study seeks to leverage the influence the main motor current, compensation motor current, and main motor loop integral error have on mechanism temperature to forecast the maximum mechanism temperature over the upcoming 6 min. A long short-term memory (LSTM) neural network predicts maximum mechanism temperature using ATMS RDR telemetry data as the input. The performance of the LSTM is compared with observed maximum mechanism temperatures by applying the LSTM coefficients to several cases. In all cases studied, the mean average error (MAE) of the forecast remained under 1.1 °C, and the correlation between forecasts and measurements remained above 0.96. These forecasts of maximum mechanism temperature are expected to be able to provide information on when the ATMS instrument should enter safe mode without needlessly losing valuable data for the ATMS flight operational team.

Identifying Causes of Short-Range Forecast Errors in Maximum Temperature during Recent Central European Heatwaves Using the ECMWF-IFS Ensemble

Abstract In the last few years, central Europe faced a number of severe, record-breaking heatwaves. Previous studies focused on predictability of heatwaves on medium-range to subseasonal time scales (5–30 days). However, also short-range (3-day) forecasts of maximum temperature (Tmax) can exhibit substantial errors even on larger spatial scales. This study investigates the causes of short-range forecast errors in Tmax over central Europe for the summers of 2015–20 using the 50-member ensemble of the operational ECMWF-IFS (ECMWF-ENS). The 3-day forecast errors, individually calculated for each ensemble member with respect to a 0–18-h control forecast, are fed into a multivariate linear regression model to study the relative importance of different error sources. Outside of heatwaves, errors in Tmax forecasts are predominantly caused by incorrectly predicted downwelling shortwave radiation, mainly due to errors in low cloud cover. During heatwaves, ECMWF-ENS exhibits a systematic underestimation of Tmax (−0.4 K), which is exacerbated under clear-sky and low wind conditions, and other error sources gain importance: the second most important error source is over- or underestimation of nocturnal temperatures in the residual layer. Additional Lagrangian trajectory analysis for the years 2018–20 (due to limited data availability) suggests a link to accumulating errors in near-surface diabatic heating of air masses associated with forecast errors in residence time over land and cloud cover. Regionally, other physical processes can be of dominant importance during heatwaves. Coastal regions are influenced by errors in near-surface wind whereas errors in soil moisture are more important in southeastern parts of central Europe.

Forward variable selection for random forest models

Random forest is a popular prediction approach for handling high dimensional covariates. However, it often becomes infeasible to interpret the obtained high dimensional and non-parametric model. Aiming for an interpretable predictive model, we develop a forward variable selection method using the continuous ranked probability score (CRPS) as the loss function. eOur stepwise procedure selects at each step a variable that minimizes the CRPS risk and a stopping criterion for selection is designed based on an estimation of the CRPS risk difference of two consecutive steps. We provide mathematical motivation for our method by proving that in a population sense, the method attains the optimal set. In a simulation study, we compare the performance of our method with an existing variable selection method, for different sample sizes and correlation strength of covariates. Our method is observed to have a much lower false positive rate. We also demonstrate an application of our method to statistical post-processing of daily maximum temperature forecasts in the Netherlands. Our method selects about 10% covariates while retaining the same predictive power.

A novel ensemble learning for post-processing of NWP Model's next-day maximum air temperature forecast in summer using deep learning and statistical approaches

A reliable and accurate extreme air temperature in summer is necessary to prepare for and respond to thermal disasters such as heatstroke and power outages. The numerical weather prediction (NWP) model is commonly used to forecast air temperature using dynamic mechanisms. Because of its high uncertainty from coarse spatial resolution and unstable parameterization, however, it requires post-processing. Recent studies have proposed advanced post-processing methods using machine learning and deep learning techniques. This study compared various individual post-processing models—multi-linear regression (MLR), support vector regression (SVR), gated recurrent units (GRU), and convolutional neural network (CNN). It also proposed a novel multi-model ensemble (MMESS) that aggregates individual post-processing models based on the skill score (SS) for the Local Data Assimilation and Prediction System (LDAPS, a local NWP model over Korea) model's next-day maximum air temperature (Tmax) forecast data in two different domains: South Korea and Seoul. The pressure and surface data of the present-day analysis and next-day forecast fields of LDAPS were used as input variables. As a result of hindcast validation, CNN showed good overall performance (root mean square error (RMSE) of 1.41 °C in South Korea and 1.50 °C in Seoul) among individual models. We found that CNN demonstrated lower RMSE (1.17–1.58 °C) than other post-processing models (1.43–2.17 °C) at stations where the bias of LDAPS changes, using surrounding spatial information. The proposed MMESS exhibited more reliable, robust results than the individual models did. A further comparison to the simple average ensemble and the constrained linear squares-based MMEsupported the proposed MMESS as a more suitable ensemble method for next-day Tmax forecast, considering the relative significance of the individual models.

Вплив категорій землекористування на прогноз WRF та його точність

In this research analysis of possibilities and methods of correction or reclassification of land use and land cover data, obtained modeling effects of using more accurate land use and land cover data and estimations of improvements of surface meteorological parameters simulations by mesoscale numerical atmospheric prediction (NWP) model. For the Kyiv region area, it is shown that the default WRF land use and land cover datasets are not ideal and contain various types of inaccuracies. Thus, compared to USGS, MODIS IGBP is characterized by a better spatial resolution, more correctly describing the urban built-up territories, forest areas and the Dnipro riverbed, but MODIS IGBP does not display small settlements. Analysis of other researches had shown that most of the reviewed studies are based on the case study approach, namely on an analysis of a small number of specific simulations. The disadvantage of this approach is that the sample size from which conclusions can be drawn is insignificant, so it is useful to compare similar studies for the consistency of the obtained results. The most consistent conclusions regarding the simulation improvement by the land use and land cover modification in the WRF model are related to the parameters of air temperature and humidity. Thus, the usage of more accurate and up-to-date land use and land cover databases in WRF simulations made it possible to reduce bias of forecasts and improved the accuracy of minimum and maximum daily temperature forecasts. The most significant improvement in the performed forecasts concerns the night temperature. The second parameter that were most often able to predict better is relative humidity. Improvement of simulated wind parameters using another land use and land cover dataset in the WRF model may be possible in urban areas and near the coast, especially as improving description of breeze circulation by the model. Assessments regarding the possibility of improving WRF precipitation simulation by land use and land cover dataset modification are not definite, although some works have noted the influence of the heat island of urbanized areas on redistribution of precipitation. On the one hand, in most of the analyzed scientific works, land use and land cover changes did not allow to significantly improve precipitation forecasts using the WRF model. On the other hand, it can be assumed that the physical and geographical conditions of a simulation domain as well as available moisture in the atmosphere is playing a more significant role in precipitation modeling than land use and land cover.

Exploratory Data Analysis of Accuracy of US Weather Forecastes

How reliable are the weather forecasts? Based on data collected from onehundred and thirteen cities in the United States over 38 months on threevariables, maximum temperature, minimum temperature and precipitation,accuracy of the weather forecasts was examined. The same day forecast hasbeen extremely accurate, especially for the maximum temperature, whilethe forecast errors and variability increase as forecasts go further out indays. Some cities have larger or smaller forecast errors than the others.For long-term weather forecasts, the maximum and minimum temperatureforecast errors has decreasing correlations overtime, respectively; However,the correlation between maximum and minimum temperature forecast errorsis positive and increasing overtime. The 7-days forecast errors of precipitationare pretty accurate.

Making the Output of Seasonal Climate Models More Palatable to Agriculture: A Copula-Based Postprocessing Method

AbstractSeasonal climate forecasts from raw climate models at coarse grids are often biased and statistically unreliable for credible crop prediction at the farm scale. We develop a copula-based postprocessing (CPP) method to overcome this mismatch problem. The CPP forecasts are ensemble based and are generated from the predictive distribution conditioned on raw climate forecasts. CPP performs univariate postprocessing procedures at each station, lead time, and variable separately and then applies the Schaake shuffle to reorder ensemble sequence for a more realistic spatial, temporal, and cross-variable dependence structure. The use of copulas makes CPP free of strong distributional assumptions and flexible enough to describe complex dependence structures. In a case study, we apply CPP to postprocess rainfall, minimum temperature, maximum temperature, and radiation forecasts at a monthly level from the Australian Community Climate and Earth-System Simulator Seasonal model (ACCESS-S) to three representative stations in Australia. We evaluate forecast skill at lead times of 0–5 months on a cross-validation theme in the context of both univariate and multivariate forecast verification. When compared with forecasts that use climatological values as the predictor, the CPP forecast has positive skills, although the skills diminish with increasing lead times and finally become comparable at long lead times. When compared with the bias-corrected forecasts and the quantile-mapped forecasts, the CPP forecast is the overall best, with the smallest bias and greatest univariate forecast skill. As a result of the skill gain from univariate forecasts and the effect of the Schaake shuffle, CPP leads to the most skillful multivariate forecast as well. Further results investigate whether using ensemble mean or additional predictors can enhance forecast skill for CPP.

Quantifying the Uncertainties in Data-Driven Models for Reservoir Inflow Prediction

Reservoir inflow prediction is subject to high uncertainties in data-driven modelling. In this study, a decomposition scheme is proposed to evaluate the individual and combined contributions of uncertainties from input sets and data-driven models to the total predictive uncertainty. Six variables (i.e., inflow (Q), precipitation (P), relative humidity (H), minimum temperature (Tmin), maximum temperature (Tmax) and precipitation forecast (F)), and three data-driven models (i.e., artificial neural network (ANN), support vector machine (SVM), and adaptive neuro fuzzy inference systems (ANFIS)) are used to produce an ensemble of 10-day inflow forecast for Huanren reservoir in China, and the analysis of variance (ANOVA) method is employed to decompose the uncertainty. The ensemble forecast results show that when the three variables, i.e., Q, P and F, are used only, the predictive accuracy of the data-driven models is very high and the addition of the other three variables, i. e., H, Tmin and Tmax, can slightly improve the predictive accuracy. The decomposition results indicate that the input set is the dominant source of uncertainty, the contribution of the data-driven model is limited and has a strong seasonal variation: larger in winter and summer, smaller in spring and autumn. Most importantly, the interactive contribution of the input set and the data-driven model to the total predictive uncertainty is very high and is more significant than the individual contribution from the model itself, implying that the combined effects of the input set and the data-driven model should be carefully considered in the modelling process.

On the Joint Calibration of Multivariate Seasonal Climate Forecasts from GCMs

Abstract Multivariate seasonal climate forecasts are increasingly required for quantitative modeling in support of natural resources management and agriculture. GCM forecasts typically require postprocessing to reduce biases and improve reliability; however, current seasonal postprocessing methods often ignore multivariate dependence. In low-dimensional settings, fully parametric methods may sufficiently model intervariable covariance. On the other hand, empirical ensemble reordering techniques can inject desired multivariate dependence in ensembles from template data after univariate postprocessing. To investigate the best approach for seasonal forecasting, this study develops and tests several strategies for calibrating seasonal GCM forecasts of rainfall, minimum temperature, and maximum temperature with intervariable dependence: 1) simultaneous calibration of multiple climate variables using the Bayesian joint probability modeling approach; 2) univariate BJP calibration coupled with an ensemble reordering method (the Schaake shuffle); and 3) transformation-based quantile mapping, which borrows intervariable dependence from the raw forecasts. Applied to Australian seasonal forecasts from the ECMWF System4 model, univariate calibration paired with empirical ensemble reordering performs best in terms of univariate and multivariate forecast verification metrics, including the energy and variogram scores. However, the performance of empirical ensemble reordering using the Schaake shuffle is influenced by the selection of historical data in constructing a dependence template. Direct multivariate calibration is the second-best method, with its far superior performance in in-sample testing vanishing in cross validation, likely because of insufficient data relative to the number of parameters. The continued development of multivariate forecast calibration methods will support the uptake of seasonal climate forecasts in complex application domains such as agriculture and hydrology.

<i>Corrigendum to</i>: ACCESS-S1: The new Bureau of Meteorology multi-week to seasonal prediction system

ACCESS-S1 will be the next version of the Australian Bureau of Meteorology's seasonal prediction system, due to become operational in early 2018. The multiweek and seasonal performance of ACCESS-S1 has been evaluated based on a 23-year hindcast set and compared to the current operational system, POAMA. The system has considerable enhancements compared to POAMA, including higher vertical and horizontal resolution of the component models and state-ofthe-art physics parameterisation schemes. ACCESS-S1 is based on the UK Met Office GloSea5-GC2 seasonal prediction system, but has enhancements to the ensemble generation strategy to make it appropriate for multi-week forecasting, and a larger ensemble size. ACCESS-S1 has markedly reduced biases in the mean state of the climate, both globally and over Australia, compared to POAMA. ACCESS-S1 also better predicts the early stages of the development of the El Niño Southern Oscillation (through the predictability barrier) and the Indian Ocean Dipole, as well as multi-week variations of the Southern Annular Mode and the Madden-Julian Oscillation — all important drivers of Australian climate variability. There is an overall improvement in the skill of the forecasts of rainfall, maximum temperature (Tmax) and minimum temperature (Tmin) over Australia on multi-week timescales compared to POAMA. On seasonal timescales the differences between the two systems are generally less marked. ACCESS-S1 has improved seasonal forecasts over Australia for the austral spring season compared to POAMA, with particularly good forecast reliability for rainfall and Tmax. However, forecasts of seasonal mean Tmax are noticeably less skilful over eastern Australia for forecasts of late autumn and winter compared to POAMA. The study has identified scope for improvement of ACCESS-S in the future, particularly 1) reducing rainfall errors in the Indian Ocean and Maritime Continent regions, and 2) initialising the land surface with realistic soil moisture rather than climatology. The latter impacts negatively on the skill of the temperature forecasts over eastern Australia and is being addressed in the next version of the system, ACCESS-S2.

Bias correction of maximum temperature forecasts over India during March–May 2017

In recent times, instances of intense heat waves have increased over the Indian subcontinent. This increase in temperature has an adverse effect on human health and the economy. Over India, such high temperatures are usually seen during the months of March–May (summer). For weather forecasters, it is a challenging job to accurately predict the timing and intensity of this anomalous high temperature. The difficulty in the accurate prediction of weather is increased because of the presence of systematic biases in the models. These biases are present because of improper parameterizations or model physics. For increasing the reliability or accuracy of a forecast it is essential to remove these biases by using a process called post-processing. In this study the biases in the surface temperature maximum are corrected using two methods, namely, the moving average and the decaying average. One of the main advantages of both the methods is that they do not require a large amount of past data for calibration and they take into account the most recent behaviour of the forecasting system. Verification, for maximum surface temperature during March–May 2017, was carried out in order to decide upon the method giving the best temperature forecast. It was found that both the bias correction methods lead to a decrease in the mean error in maximum surface temperature (Tmax). However, the decaying average method showed a higher decrease in the mean error. Scores obtained from a contingency table like POD, FAR and PSS, showed that for Tmax, the decaying average method outperforms the forecasts, i.e., raw and moving average in terms of having high POD and PSS and a low FAR.

Bias Correction of Short-Range Ensemble Forecasts of Daily Maximum Temperature Using Decaying Average

In this study, we assessed the performance of the decaying average bias correction method in removing the systematic error in daily maximum temperature (dTmax) ensemble forecasts. We applied the technique to a short-range high-resolution limited-area ensemble prediction system of the Korea Meteorological Administration, which shows under-predictive and under-dispersive characteristics for dTmax. The bias correction was applied to the grid of the model using spatial interpolation of the decaying average bias from surrounding reference points. The method was verified by evaluating the accuracy of the ensemble mean, spread-skill relationship, and the performance of the probabilistic forecasts. The results showed that the decaying average technique minimized the systematic error in the ensemble mean and improved the performance of the probabilistic forecasts. The overall mean absolute error of the ensemble mean was lowered from 2.2 to 1.2 ∘C and the root-mean-square error from 2.5 to 1.6 ∘C. The continuous ranked probability score decreased from 1.9 to 1.0 ∘C. The reliability of three dichotomous events also improved and the Brier skill scores increased. However, the bias correction only slightly affected the ensemble spread, and the system remained under-dispersive.

The Seasonality and Regionality of MJO Impacts on North American Temperature

AbstractIt is widely accepted that the Madden‐Julian Oscillation's (MJO) influence on North American temperature is strongest in winter. A growing body of literature demonstrates that the MJO also influences North American weather in other seasons. Here we use observations to investigate the seasonality and regionality of the MJO's impact on weather station daily maximum air temperature over North America (NA). Consistent with previous work, we find the strongest MJO signal in temperatures over eastern NA and Alaska during winter. However, the peak MJO signals over much of central NA, western NA, and south Texas occur outside of winter. We investigate how this translates to forecast skill and conduct leave‐one‐out cross‐validated empirical forecasts of maximum surface air temperature using the phase of the MJO and lead time as predictors. Our results suggest the potential for more skillful long‐range forecasts of weather over NA during spring, summer, and fall.

A Prediction Scheme for Daily Maximum and Minimum Temperature Forecasts Using Recurrent Neural Network and Rough set

A new nonlinear objective prediction scheme has been developed for predicting 24h daily maximum and minimum temperature forecasts at 14 stations in Guangxi, China during Jan, 2015-Jun, 2018 using Recurrent Neural Network (RNN) and based on the daily average, maximum, minimum temperature and precipitation data. Taking the climatology and persistence predictors as primary factors, the conditional attribute reduction method of rough set theory is adopted. By eliminating the unrelated attributes, the predictors direct correlated with the predictand (maximum and minimum temperature) are taken as the RNN model input by means of attribute reduction. This new scheme is validated with 24h short-range forecasts spanning Jan to Jun, 2018. Using identical predictors and sample cases, predictions of the RNN model are compared with the stepwise regression method, and results show that the former is more accurate. The mean absolute errors of RNN at 14 stations in Guangxi are lower than those of the stepwise regression method. The mean forecast accuracy with absolute errors being less than 2°C (1°C) of RNN is higher than that of the stepwise regression method. Moreover, the number of forecast errors larger than 2°C and the system deviation of daily maximum (minimum) temperature prediction are significantly reduced by RNN model, indicating a potentially better operational weather prediction tool.

中央台指导预报在塔城地区最高温度预报中的对比检验及评估

中央台指导预报在塔城地区最高温度预报中的对比检验及评估

Assessing the suitability of statistical downscaling approaches for seasonal forecasting in Senegal

AbstractThis work tests the suitability of statistical downscaling (SD) approaches to generate local seasonal forecasts of daily maximum temperature and precipitation for a set of selected stations in Senegal for the July–August–September season during the period 1979–2000. Two‐month lead raw daily maximum temperature and precipitation from the five models included in the ENSEMBLES seasonal hindcast are compared against the corresponding downscaled predictions, which are obtained by applying the analog technique based on two different types of predictors: the direct surface variables and a combination of appropriate upper‐air variables. Beyond correcting the large biases of the low‐resolution raw model outputs, SD is found to add noteworthy value in terms of forecast association (as measured by interannual correlation), providing thus suitable (i.e. calibrated) predictions at the local‐scale needed for practical applications, which means a clear advantage for the end‐users of seasonal forecasts over the area of study. Moreover, a recommendation on the adequacy of surface (large‐scale) predictors for SD of maximum temperature (precipitation) is also given.