@article{VormoorSkaugen2013,
  author    = {Vormoor, Klaus Josef and Skaugen, Thomas},
  title     = {Temporal disaggregation of daily temperature and precipitation grid data for Norway},
  series = {Journal of hydrometeorology},
  volume    = {14},
  journal   = {Journal of hydrometeorology},
  number    = {3},
  publisher = {American Meteorological Soc.},
  address   = {Boston},
  issn      = {1525-755X},
  doi       = {10.1175/JHM-D-12-0139.1},
  pages     = {989 -- 999},
  year      = {2013},
  abstract  = {This paper presents a simple approach for the temporal disaggregation from daily to 3-hourly observed gridded temperature and precipitation (1 x 1km(2)) on the national scale. The intended use of the disaggregated 3-hourly data is to recalibrate the hydrological model currently used by the Norwegian Water Resources and Energy Directorate (NVE) for daily flood forecasting. By adapting the hydrological model to a 3-hourly temporal scale, the flood forecasting can benefit from available meteorological forecasts with finer temporal resolution and can better represent critical events of short duration and at small spatial scales. By consulting the temporal patterns of a High-Resolution Limited-Area Model (HIRLAM) hindcast series for northern Europe with an hourly temporal and a 0.1 degrees spatial resolution, existing daily 1 x 1km(2) grids for temperature and precipitation covering all of Norway (the seNorge data) were disaggregated into 3-hourly values for the time period September 1957 to December 2010. For the period 2000-05, the disaggregated 3-hourly temperature and precipitation data are validated against observed values from five meteorological stations and against 3-hourly data from the HIRLAM hindcast and daily seNorge data simply split into eight fractions. The results show that the disaggregated data perform best with anomaly correlation coefficients between 0.89 and 0.92 for temperature. With regard to precipitation, the disaggregated data also provide the highest correlations and the lowest errors. In addition, the disaggregated data prove to be best in estimating intervals without precipitation and tend to be most appropriate in estimating extreme precipitation with low occurrence probability (<20\%).},
  language  = {en}
}