Mapping rice fields in Sacramento Valley, CA using Google Earth Engine and Sentinel-2 Multitemporal Data

In this project, I utilized Google Earth Engine and Remote Sensing techniques to map rice fields in Sacramento Valley, CA, using multitemporal Sentinel-2 data. The primary focus was on flooding patterns in rice fields over specific time periods, which allowed for tracking changes and calculating the total rice field area.

I developed a reusable script that begins with creating mosaics of multiple Sentinel-2 scenes to cover the entire area of interest, incorporating cloud masking for improved accuracy. The mosaic was then segmented, dividing the image into clusters based on spectral values to distinguish different crop types. This process allowed for the extraction and merging of multitemporal data into a single file that contained all detected rice fields during a reference period, such as May to July 2024.

The final data was available in both raster and vector formats, with detailed statistics generated and exported in user-friendly formats like .tif, .shp, and .xls for further analysis. This approach not only improved the accuracy of rice field mapping but also allowed for continuous monitoring using Sentinel-2 imagery. The project also employed cluster classification methods and the application of effective cloud masks to minimize errors, ensuring that the output was reliable for subsequent GIS applications and decision-making.