Determining and mapping the beginning of summer crops cycle through satellite imagery and rainfall data
Abstract
Monitoring agricultural coupled with studies of weather variability is extremely important because they determine the best time which should be sown as well as periods of higher water requirement of the plant. The use of remote sensing for agricultural monitoring in addition to having less subjectivity in information, when coupled with meteorological data enables the monitoring of crops from planting to harvest. The main goal of this study was mapping the beginning of the soybean cycle in Paraná state, Brazil, through dekadal composition from SPOT Vegetation for 2005/2006, 2006/2007 and 2007/2008 crop season. We used a supervised digital classification method with Parallelepiped algorithm in multitemporal RGB image composites to generate soybean crop masks for each 10-day composition. The accuracy assessment for soybean masks reached Kappa index values of 0.70, 0.75 and 0.77 and overall accuracy 0.91, 0.91 and 0.93. We also integrated data from SIMEPAR network meteorological ground stations to the ECMWF atmospheric model data, during the beginning of the 2005/06, 2006/07 e 2007/08 crop season. We interpolated these data using the inverse square of the distance achieving good results, since the actual values were significant when compared to the interpolated values in the data validation, when evaluated by the agreement index 'd' of Willmot it ranged between 0.85 to 0.94, showing a good spatial distribution of rainfall in this region. Then we crossed the crop season masks with the dekadal rainfall maps, in order to determine and to map the beginning of the crop vegetative cycle. The association between spectral and rainfall data have allowed the identification and mapping of the differences in the start of the crop cycle. Therefore we noticed there was a delay at the beginning of the crop cycle within areas where the sowing occurred before the rainy season in the 2007/2008 crop season. Besides, we found inside of a specific region that there is more than one sowing date for crop season.
- Publication:
-
AGU Fall Meeting Abstracts
- Pub Date:
- December 2013
- Bibcode:
- 2013AGUFM.B41A0370F
- Keywords:
-
- 0402 BIOGEOSCIENCES Agricultural systems;
- 0480 BIOGEOSCIENCES Remote sensing