CR 57:1-9 (2013)  -  DOI: https://doi.org/10.3354/cr01164

Validation of ELPIS 1980-2010 baseline scenarios using the observed European Climate Assessment data set

Mikhail A. Semenov1,*, Scott Pilkington-Bennett2, Pierluigi Calanca3

1Computational and Systems Biology Department, and 2Biological Chemistry and Crop Protection, Rothamsted Research, Harpenden, Hertfordshire AL5 2JQ, UK
3Agroscope Reckenholz-Tänikon, Natural Resources and Agriculture, 8046 Zurich, Switzerland

ABSTRACT: Local-scale daily climate scenarios are required for assessment of climate change impacts. ELPIS is a repository of local-scale climate scenarios for Europe, which are based on the LARS-WG weather generator and future projections from 2 multi-model ensembles, CMIP3 and EU-ENSEMBLES. In ELPIS, the site parameters for the 1980-2010 baseline scenarios were estimated by LARS-WG using daily weather from the European Crop Growth Monitoring System (CGMS) used in many European agricultural assessment studies. The objective of this paper was to compare ELPIS baseline scenarios with observed daily weather obtained independently from the European Climate Assessment (ECA) data set. Several statistical tests were used to compare distributions of climatic variables derived from ECA-observed daily weather and ELPIS-generated baseline scenarios. About 30% of selected sites have a difference in altitude of >50 m compared with the CGMS grid-cell altitude that was selected to represent agricultural land within a grid-cell. Differences in altitude can explain significant Kolmogorov-Smirnov test (KS-test) results for distribution of daily temperature and in t-tests for temperature monthly means, because of the well-known negative correlation between temperature and elevation. For daily precipitation, the KS-test showed little difference between generated and observed data; however, the more sensitive t-test showed significant results for the sites where altitude differences were large. Approximately 11% of sites showed small positive or negative bias in monthly solar radiation, although 86% sites showed >3 significant t-test results for monthly means. These results can be explained by differences in conversion of sunshine hours to solar radiation used in CGMS and LARS-WG. We conclude that, considering the limitations above, ELPIS baseline scenarios are suitable for agricultural impact assessments in Europe.


KEY WORDS: Climate change · Impact assessment · Downscaling · LARS-WG


Full text in pdf format 
Supplementary material  
Cite this article as: Semenov MA, Pilkington-Bennett S, Calanca P (2013) Validation of ELPIS 1980-2010 baseline scenarios using the observed European Climate Assessment data set. Clim Res 57:1-9. https://doi.org/10.3354/cr01164

Export citation
Mail this link - Contents Mailing Lists - RSS
- -