Pregled bibliografske jedinice broj: 229978
Geostatistical Methods in Producing Mean Annual Precipitation Map for Croatia
Geostatistical Methods in Producing Mean Annual Precipitation Map for Croatia // European Conference on Applied Climatology (ECAC 2002) : Abstracts Volume
Brisel: Royal Meteorological Insitute of Belgium, 2002. (poster, nije recenziran, sažetak, znanstveni)
CROSBI ID: 229978 Za ispravke kontaktirajte CROSBI podršku putem web obrasca
Naslov
Geostatistical Methods in Producing Mean Annual Precipitation Map for Croatia
Autori
Perčec Tadić, Melita ; Gajić-Čapka, Marjana ; Patarčić, Mirta
Vrsta, podvrsta i kategorija rada
Sažeci sa skupova, sažetak, znanstveni
Izvornik
European Conference on Applied Climatology (ECAC 2002) : Abstracts Volume
/ - Brisel : Royal Meteorological Insitute of Belgium, 2002
Skup
European Conference on Applied Climatology
Mjesto i datum
Bruxelles, Belgija, 12.11.2002. - 15.11.2002
Vrsta sudjelovanja
Poster
Vrsta recenzije
Nije recenziran
Ključne riječi
precipitation map; multiple linear regression; geostatistical universal kriging; DEM; geographic information system (GIS)
Sažetak
Mean annual precipitation map for Croatia has been produced for the last CLINO period 1961-1990. Data from the heterogeneously weather station network have been used. It included 643 stations. Statistical methods, namely Multiple Linear Regression (MLR) and Geostatistical Universal Kriging, have been applied to estimate a precipitation field on a regular grid. Resolution of the grid has been 700x700 m. The first method, MLR, reveals the spatial trend that exists in the precipitation data. Four variables, longitude (x), latitude (y), elevation (z) and distance from the sea (con) of the meteorological stations have been used as predictors. They were elaborated from the Digital Elevation Model (DEM) which was part of the GIS (Geographic Information System). According to this model and measured precipitation first estimate of the mean annual precipitation has been established: R (mm) = a_0 + a_1 (x (m)) + a_2 (y(m)) + a_3 (z(m)) + a_4 (con(m)). Model explains 61% of the variance in data. The next step was interpolation of the differences between measured and predicted values at the meteorological stations, so called residuals, to the regular grid. The second method, Universal Kriging, was applied for the spatial interpolation of the residual terms. It can be regarded as Residual or Detrended Kriging. Weighting factors were estimated according to semi-variogram model. Interpolated residuals and preliminary field from MLR analysis were combined to produce a final precipitation field. There were the evident errors in the estimates of precipitation amounts at higher elevations due to rapid change of height and spatial resolution of the model. Because of that the corrections have been made according to the regional vertical gradients of the annual precipitation. Mean annual precipitation distribution is characterized by minimum amounts at the most distant south Adriatic islands (Palagruza-311 mm) and maximum in the mountainous region of Gorski kotar (Lividraga-3728 mm). The greatest increase of precipitation with height occurs at the windward side of mountains along the eastern Adriatic coast due to the upward humid maritime air flow (south part of Velebit mountain and hills north-east of the Konavli field). According to this, precipitation amounts are smaller on the leeward side at the same height. In the continental part of Croatia precipitation increases from east (600-700 mm in east Slavonia) to west, reaching highest values between 1000 and 1500 mm in regions with pronounced orography (Zagorje, Samobor hills, Medvednica and Kalnik).
Izvorni jezik
Engleski
Znanstvena područja
Geologija
POVEZANOST RADA
Ustanove:
Državni hidrometeorološki zavod