An evaluation of statistical models for downscaling precipitation and their ability to capture long‐term trends

Benestad, Rasmus E.; Hanssen-Bauer, Inger; Førland, Eirik J.

doi:10.1002/joc.1421

Cited by 56 publications

(35 citation statements)

References 24 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Benestad et al (2007) highlighted that, there exists a risk that a statistical downscaling method that performs well under present climate could fail under changing climate in future, simply due to the invalidity of the above-mentioned assumption. Statistical downscaling techniques are classified under three main categories of weather classification, regression models and weather generators (Wilby et al, 2004).…”

Section: Introductionmentioning

confidence: 99%

Least square support vector and multi‐linear regression for statistically downscaling general circulation model outputs to catchment streamflows

Sachindra

Huang

Barton

et al. 2012

Intl Journal of Climatology

113

View full text Add to dashboard Cite

This study employed least square support vector machine regression (LS-SVM-R) and multi-linear regression (MLR) for statistically downscaling monthly general circulation model (GCM) outputs directly to monthly catchment streamflows. The scope of the study was limited to calibration and validation of the downscaling models. The methodology was demonstrated by its application to a streamflow site in the Grampian water supply system in northwestern Victoria, Australia. Probable predictors for the study were selected from the National Center for Environmental Prediction/National Center for Atmospheric Research (NCEP/NCAR) reanalysis data set based on the past literature and hydrology. Probable variables that displayed the best significant correlations, consistently with the streamflows over the entire period of the study and under three 20-year time slices (1950-1969, 1970-1989 and 1990-2010) were selected as potential predictors. To better capture seasonal variations of streamflows, downscaling models were developed for each calendar month. The standardized potential predictors were introduced to the LS-SVM-R and MLR models, starting from the best correlated three and then, others one by one, based on their correlations with the streamflows, until the model performance in validation was maximized. This stepwise model development enabled the identification of the optimum number of potential variables for each month. The model calibration was performed over the period 1950-1989 and validation was done for 1990-2010. LS-SVM-R model parameter optimization was achieved using simplex algorithm and leave-one-out cross-validation. The MLR models were optimized by minimizing the sum of squared errors. In both modelling techniques, validation was performed as an independent simulation. In calibration, LS-SVM-R and MLR models displayed equally good performances with a trend of under-predicting high flows. During validation, LS-SVM-R outperformed MLR, though both techniques over-predicted most of the streamflows. It was concluded that LS-SVM-R is a better technique for statistically downscaling GCM outputs to streamflows than MLR, but still MLR is a potential technique for the same task.

show abstract

Section: Introductionmentioning

confidence: 99%

Least square support vector and multi‐linear regression for statistically downscaling general circulation model outputs to catchment streamflows

Sachindra

Huang

Barton

et al. 2012

Intl Journal of Climatology

113

View full text Add to dashboard Cite

show abstract

“…A tendency of increasing precipitation in winter and spring was detected during the latter half of the twentieth century. Benestad et al (2007) compared downscaled and modelled precipitation at 27 stations in Fennoscandia for . Only a few locations exhibited trends that were statistically significant at the 5 % error level.…”

Section: Precipitation Climate In Recent Decadesmentioning

confidence: 99%

Recent Change—Atmosphere

Rutgersson

Jaagus

Schenk

et al. 2015

Regional Climate Studies

View full text Add to dashboard Cite

This chapter describes observed changes in atmospheric conditions in the Baltic Sea drainage basin over the past 200-300 years. The Baltic Sea area is relatively unique with a dense observational network covering an extended time period. Data analysis covers an early period with sparse and relatively uncertain measurements, a period with well-developed synoptic stations, and a final period with 30+ years of satellite data and sounding systems. The atmospheric circulation in the European/Atlantic sector has an important role in the regional climate of the Baltic Sea basin, especially the North Atlantic Oscillation. Warming has been observed, particularly in spring, and has been stronger in the northern regions. There has been a northward shift in storm tracks, as well as increased cyclonic activity in recent decades and an increased persistence of weather types. There are no long-term trends in annual wind statistics since the nineteenth century, but much variation at the (multi-)decadal timescale. There are also no long-term trends in precipitation, but an indication of longer precipitation periods and possibly an increased risk of extreme precipitation events.

show abstract

“…In this respect, the degree of non-stationarity between predictand and predictor has been considered by Hewitson and Crane (2006), while Benestad et al (2007) and Fan et al (2011) highlighted the difficulties in capturing longterm trends through downscaling when the GCM fails in this attempt.…”

Section: N Guyennon Et Al: Benefits From Using Combined Dynamical-smentioning

confidence: 99%

Benefits from using combined dynamical-statistical downscaling approaches – lessons from a case study in the Mediterranean region

Guyennon

Romano

Portoghese

et al. 2013

Hydrol. Earth Syst. Sci.

View full text Add to dashboard Cite

Abstract. Various downscaling techniques have been developed to bridge the scale gap between global climate models (GCMs) and finer scales required to assess hydrological impacts of climate change. Such techniques may be grouped into two downscaling approaches: the deterministic dynamical downscaling (DD) and the statistical downscaling (SD). Although SD has been traditionally seen as an alternative to DD, recent works on statistical downscaling have aimed to combine the benefits of these two approaches. The overall objective of this study is to assess whether a DD processing performed before the SD permits to obtain more suitable climate scenarios for basin scale hydrological applications starting from GCM simulations. The case study presented here focuses on the Apulia region (South East of Italy, surface area about 20 000 km 2 ), characterised by a typical Mediterranean climate; the monthly cumulated precipitation and monthly mean of daily minimum and maximum temperature distribution were examined for the period 1953-2000. The fifth-generation ECHAM model from the MaxPlanck-Institute for Meteorology was adopted as GCM. The DD was carried out with the Protheus system (ENEA), while the SD was performed through a monthly quantile-quantile correction. The SD resulted efficient in reducing the mean bias in the spatial distribution at both annual and seasonal scales, but it was not able to correct the miss-modelled nonstationary components of the GCM dynamics. The DD provided a partial correction by enhancing the spatial heterogeneity of trends and the long-term time evolution predicted by the GCM. The best results were obtained through the combination of both DD and SD approaches.

show abstract

An evaluation of statistical models for downscaling precipitation and their ability to capture long‐term trends

Cited by 56 publications

References 24 publications

Least square support vector and multi‐linear regression for statistically downscaling general circulation model outputs to catchment streamflows

Least square support vector and multi‐linear regression for statistically downscaling general circulation model outputs to catchment streamflows

Recent Change—Atmosphere

Benefits from using combined dynamical-statistical downscaling approaches – lessons from a case study in the Mediterranean region

Contact Info

Product

Resources

About