Keywords

climate change, weather data, crop growth modelling, europe

Start Date

1-7-2012 12:00 AM

Abstract

Global Circulation Models (GCM) estimate future climate under scenarios of greenhouse gases emissions. Such estimates include several meteorological parameters but the two direct outputs are air temperature at earth surface and precipitation. The estimates are spatially downscaled using different methodologies, but it is accepted that such data require further processing for use with simulation models. Daily values of solar radiation, wind, air humidity, and at times rainfall may have absolute values which are not realistic, and/or the daily record of data may prove not to be consistent across meteorological variables. The final problem is related to the fact that GCM estimate the dynamics of climate, providing one instance of data per date in time series. Typically, crop models are deterministic and run in a stochastic fashion, hence requiring multiple years of weather data representing each time horizon of interest. Furthermore, if the time horizons of interest are very close (e.g. 2020 and 2030), sampling without overlap GCM outputs creates instability in means which may even show, in specific cases, apparent inversions of trends, creating artifacts also in the simulation via impact models. This paper presents a data base of daily weather data, with EU27 coverage at a 25 km grid, derived from the ENSEMBLES downscaling of the global circulation models HadCM3 and ECHAM5 realizations of the IPCC A1B emission scenario, in which solar radiation, wind and relative air humidity where estimated or collected from historical series, and derived variables reference evapotranspiration and vapour pressure deficit were estimated from other variables, ensuring consistency within daily records. Synthetic time series data were also generated using the weather generator ClimGen. All data are made available via web services in a data portal that also contain link to a reference domain ontology, and meta-information, according to the specifications for the semantic web publishing of the EU framework project TaToo.

COinS
 
Jul 1st, 12:00 AM

A EU27 Database of Daily Weather Data Derived from Climate Change Scenarios for Use with Crop Simulation Models

Global Circulation Models (GCM) estimate future climate under scenarios of greenhouse gases emissions. Such estimates include several meteorological parameters but the two direct outputs are air temperature at earth surface and precipitation. The estimates are spatially downscaled using different methodologies, but it is accepted that such data require further processing for use with simulation models. Daily values of solar radiation, wind, air humidity, and at times rainfall may have absolute values which are not realistic, and/or the daily record of data may prove not to be consistent across meteorological variables. The final problem is related to the fact that GCM estimate the dynamics of climate, providing one instance of data per date in time series. Typically, crop models are deterministic and run in a stochastic fashion, hence requiring multiple years of weather data representing each time horizon of interest. Furthermore, if the time horizons of interest are very close (e.g. 2020 and 2030), sampling without overlap GCM outputs creates instability in means which may even show, in specific cases, apparent inversions of trends, creating artifacts also in the simulation via impact models. This paper presents a data base of daily weather data, with EU27 coverage at a 25 km grid, derived from the ENSEMBLES downscaling of the global circulation models HadCM3 and ECHAM5 realizations of the IPCC A1B emission scenario, in which solar radiation, wind and relative air humidity where estimated or collected from historical series, and derived variables reference evapotranspiration and vapour pressure deficit were estimated from other variables, ensuring consistency within daily records. Synthetic time series data were also generated using the weather generator ClimGen. All data are made available via web services in a data portal that also contain link to a reference domain ontology, and meta-information, according to the specifications for the semantic web publishing of the EU framework project TaToo.