4.6 Article

Mapping spatial-temporal nationwide soybean planting area in Argentina using Google Earth Engine

Journal

INTERNATIONAL JOURNAL OF REMOTE SENSING
Volume 43, Issue 5, Pages 1725-1749

Publisher

TAYLOR & FRANCIS LTD
DOI: 10.1080/01431161.2022.2049913

Keywords

Soybean; Argentina; Random Forest; Google Earth Engine

Funding

  1. National Key Research and Development Program of China [2017YFC0209700, 2020YFC1807501]
  2. National Innovation and Entrepreneurship Training Program for College Students [201910335060]

Ask authors/readers for more resources

This study successfully extracted and analyzed the nationwide soybean planting areas in Argentina from 2016 to 2019 using the Google Earth Engine and pixel-based machine learning method random forest. The results showed the importance of NDVI and NIR features in the classification. This research provides an effective method for accurately and rapidly retrieving the soybean planting area.
Argentina is one of the three most crucial soybean-producing countries in the world. Timely and accurate monitoring of the soybean planting area and its changes across the country is extremely important for global food security. This study effectively extracted the spatial-temporal nationwide soybean planting areas in Argentina from 2016 to 2019 based on the Google Earth Engine (GEE) and pixel-based machine learning method random forest. About 8000 soybean sampling points were gathered from the high spatial resolution imagery available on Google Earth. Comprehensive consideration of phenological period, monitoring area and imagery availability, multi-bands of Landsat 8 imagery and Normalized Difference Vegetation Index (NDVI) waverage annual temperatures range as used to map dynamics of nationwide soybean planting areas. The primary soybean planting areas were mapped between 2016 and 2019 and occupied 5.90%-6.76% of Argentina's total land area. Soybean planting areas at the provincial level explained more than 84% of the variability in most years compared with those obtained from official statistics. The spatial distribution of primary soybean planting areas changed from 2016 to 2019 resulting in the shift of area-weighted centroids. Results showed that NDVI and NIR features calculated by the first three months' composites contributed more to the classification. This study provides an effective way to retrieve the soybean planting area accurately and rapidly, especially when lacking chances for field measurement over a large spatial extent.

Authors

I am an author on this paper
Click your name to claim this paper and add it to your profile.

Reviews

Primary Rating

4.6
Not enough ratings

Secondary Ratings

Novelty
-
Significance
-
Scientific rigor
-
Rate this paper

Recommended

No Data Available
No Data Available