UNI-MB - logo
UMNIK - logo
 
(UL)
  • Modification of the sampling scheme for the implementation of the MARS project in Slovenia
    Šabić, Danijela ; Tretjak, Ana ; Orešnik, Irena
    The co-operation between the Remote Sensing Team at the Statistical Office of the Republic of Slovenia and the Joint Research Centre (JRC) at Ispra has beeninitiated in 1993 when JRC purchased the ... Landsat-TM satellite scanned datacovering the whole territory of Slovenia. In spring 1994 the implementation of Action 1 of MARS (Monitoring Agriculture with Remote Sensing) project was approved. If was agreed that in the first year the new MARS methodology of collecting agrostatistical data will be applied on one third of Slovenia. 22 agricultural most intensive municipalities of the northeastern area were selected. One full Landsat-TM scene and four juxtaposed quarter scenes were georeferenced with a positional error less than one pixel, i.e. 30 m. The mosaic of Slovenia with dominant reflectance values of fundamental land use classes in scale 1 : 75,000 was printed and boundariesof 6 strata were drawn and digitised. The delineating criteria for each stratum was the intensity of agricultural land. The area of the 22 municipalities was grided into blocks of 10 km x 10 km size, and 25 x 25 segments of size 400 m x 400 m were drawn in each block. Of the 625 segments per block a number of segments defined in advance was randomly selected, thus covering a defined sampling rate per stratum. The segments were transposed on aerial photos and a ground survey was performed. Data on identified crops on fields for each segment were scanned and extrapolated to the acreage of each stratum. In addition the standard error of the estimated acreage for each cropwas computed. The obtained results will be compared with the official statistical data and the obtained differences analysed with the help of classified satellite scanned data of the same area. In addition the segments were used for the yield survey. A grid of 10 by 10 nodes was superimposed overone segment and 4 nodes were randomly selected. This template was used on the rest of the 185 segments. The farmers cultivating the fields identified bythe 4 nodes were asked to participate in the yield survey. The selected farmers have to fill-in an enquiry, giving data on the acreage of the land useof their farm and data on yield of selected crops. The survey will be concluded at the end of October. The obtained data will be extrapolated to thearea of each stratum.
    Vrsta gradiva - prispevek na konferenci
    Leto - 1996
    Jezik - angleški
    COBISS.SI-ID - 23515229