High Resolution Mapping of Cropping Cycles by Fusion of Landsat and MODIS Data

Abstract

Multiple cropping, a common practice of intensive agriculture that grows crops multiple times in the agricultural land in one growing season, is an effective way to fulfill the food demand given limited cropland areas. Deriving cropping cycles from satellite data provides the spatial distribution of cropping intensities that allows for monitoring of the multiple cropping activities over large areas. Although efforts have been made to map cropping cycles at 500 m or coarser resolution, producing cropping cycle maps at high resolution remain challenging because data from single satellite sensor do not provide sufficient spatiotemporal observations. In this paper, we generate dense time series of satellite data at 30 m resolution by fusion of Landsat and MODIS data, and derive the cropping cycles from the fused time series data. The method achieves overall accuracies of 92.5% and 89.2%, respectively, for two typical regions of multiple cropping in China using samples identified based on satellite time series data, and an overall accuracy of 81.2% for four subregions using all samples identified based on multi-temporal high resolution images. The mapped crop cycles show to be reasonable geographically and agree with the national census data. The fusion approach provides a feasible way to map cropping cycles at 30 m resolution and enables improved depiction of the spatial distribution of multiple cropping.

Phenology-Based Rice Paddy Mapping Using Multi-Source Satellite Imagery and a Fusion Algorithm Applied to the Poyang Lake Plain, Southern China

“…However, previous studies noted that Landsat ETM+ and OLI often overlook some rice paddy regions, mainly due to restrictions of data availability and quality [13,46]. This study tested the fusion method in southern China and confirmed that the fusion of MODIS and Landsat images was a feasible strategy for adding more available data (Figures 7 and 8) to improve the discernment capability of phenological information for rice paddy classification [26]. Second, pixel-and phenology-based algorithms were applied for mapping rice paddies.…”

“…Compared with previous algorithms used in areas with a single-cropping system [14], our approach could also be applied to improve the mapping of rice paddy cropping intensity in areas with complex cropping systems in southern China. Masking other land-cover types has proven essential for mapping rice paddies and cropping intensity [13], and improved mapping of rice paddy planting areas at high resolutions can improve mapping of cropping intensity [26]. Our study found that the multiple cropping indices were approximately 150%.…”

“…With the Southern China is one of the most famous rice production bases over the world. However, accurate mapping of the rice paddy fields at fine resolutions in southern China is still a challenge [13,26]. First, the cropland-filled are in southern China is quite small, and the landscape is heterogeneous.…”

Spatiotemporal characteristics of white mold and impacts on yield in soybean fields in South Dakota

“…ESTARFM was selected because of its advantages: it works better in heterogeneous regions such as the croplands in our study area, it improves the prediction with the use of several bands ( Table 1 shows the bands that were combined in the algorithm) in selecting similar pixels, and it uses spectral similarities correlation coefficients between Landsat and MODIS in the weight calculation of similar pixels. Furthermore, this algorithm has outperformed many others (Emelyanova et al 2013;Li et al 2017;Wu et al 2016). Details about the ESTARFM algorithm are provided by Zhu et al (2010).…”

Cropland Mapping Using Fusion of Multi-Sensor Data in a Complex Urban/Peri-Urban Area

“…The weighted function based methods include the Spatial and Temporal Adaptive Reflectance Fusion Model (STARFM) and the Enhanced STARFM (ESTARFM) , which assume that no land cover type changes occur between the reference and prediction dates [43,44]. While this assumption limits the performance of weight function based algorithms in heterogeneous landscapes where rapid, abrupt changes occur, they are popular since they require no auxiliary data as inputs and are robust enough to predict pixels with changes in biophysical attributes [44][45][46]. In remote sensing, indices enhance spectral information and class separability and are, therefore, an essential basis for the estimation of the biophysical characteristics of land cover, such as vegetation vigor [44,46].…”