Abstract:

To reduce the consumption of human and material resources, and to improve the efficiency of identifying the green-up period of alfalfa and the accuracy of yield estimation through satellite remote sensing, using the data from the European Space Agency Sentinel-2, and the Taigemu alfalfa planting base in Tumd Left Banner, Inner Mongolia was used as the research area to study the usability and accuracy of satellite remote sensing in identifying the green-up period and estimating yields on the basis of comparing the actual situation with manual observations. The identification and extraction of green-up was carried out by using the sum of the differences of the normalized vegetation index (NDVI). The yield estimation of alfalfa was carried out using the NDVI power function model. The results showed that the remote sensing methods for the five green-up periods (include seedling stage) in 2024 all met the common green-up period conditions stipulated in the observation norms and could make accurate predictions. In all four mowing yield estimates, the yield could be predicted efficiently, and the prediction accuracy rate for the total annual yield was higher than that of a single prediction, reaching as high as 88.4%. It indicates that the methods used in this study for identifying the green-up period and estimating the yield can be extended and applied to artificial alfalfa plots with an environment similar to that of the experimental area, improving the efficiency of the investigation work, better serving relevant enterprises and farmers, and providing decision-making basis for government departments and related units.

Key words: artificial alfalfa cultivation, the green-up period, yield estimation, satellite remote sensing, normalized difference vegetation index (NDVI), Hohhot