Konferenzbeitrag
Determining the Climatologically Suitable Areas for Wheat Production Using MODIS-NDVI in Mashhad, Iran
Vorschaubild nicht verfügbar
Volltext URI
Dokumententyp
Text/Conference Paper
Zusatzinformation
Datum
2008
Autor:innen
Zeitschriftentitel
ISSN der Zeitschrift
Bandtitel
Verlag
Shaker Verlag
Zusammenfassung
One of the most important environmental matters is adaptability of crops to climate of their cultivation areas.. Geoinformatics data and tools such as satellite images and GIS can be used to develop models for determining the climatologically suitable areas for crops. In this study we tried to find a model to determine potential wheat cultivation area in Mashhad (North East of Iran). Vegetative vigour or "greenness" of wheat considered as an appropriate index to measure water availability and deficiency stress and also plant health, plant density and quality. The index is called Normalized Difference Vegetation Index (NDVI). In this study MODIS-NDVI values were compared with climatological parameters to assess the relations between vegetative vigour and climatological parameters. The NDVI values for three selected wheat farms in Mashhad area were calculated using MODIS images for 2003 and 2004 growing seasons. The data of four climatological parameters including air temperature, precipitation, relative humidity, and sunshine hours were also collected from the nearest weather stations. Then a multi-regression statistical analysis was performed to find the relation between wheat NDVI and climatological parameters in the study area. Pertaining statistical methods including Mixed, and Stepwise (Forward and Backward) were used in the analysis. Scattering matrix was used to determine the data scattering of the models and NDVI values for comparison. The results showed that backward method was more appropriate than the other two methods for predicting NDVI values of the study area. After finalizing this model the results were statistically tested using 20% of the samples for the test purpose and the remaining 80%, for running the model. The results showed that there was no significant difference between Backward, Testing Backward and Training Backward models. The results from the latter method showed that the NDVI of the pixels could be estimated for 79% of the cases. It can be stated that the rest of NDVI values could be affected by other environmental parameters such as soil type and characteristics, topographical conditions, agronomical practices, plant diseases and other unknown factors. Finally, some maps were developed showing the potential wheat farming in the area according to the model results.