We tested if a new vegetation index, which combines the chlorophyll index and a measure of solar radiation (Σ (CI×I) ) , could estimate the dry matter weight (DMW) of soybeans by proximal remote sensing. In each year of the three-year experiment, we made two kinds of soybean communities which had different row spaces. We judged the usefulness of Σ (CI×I) by comparing it to both the ratio vegetation index (RVI) and an index combining the normalized differential vegetation index and solar radiation (Σ (NDVI×I) ). The values of Σ (CI×I) and Σ (NDVI×I) were found by the following methods. We measured CI and NDVI 3–6 times over the soybean growth period using a handy-type-spectral-reflectance-measuring-device and estimated a daily value of the CI and NDVI by using spline-smoothing regression method. Then, we multiplied CI or NDVI by solar radiation (MJ/m
2/day) , and summed this value over a period from the sowing date to the DMW measurement date of the soybeans. We measured DMWs of soybeans 3 - 6 times during the growth periods, and we examined the relationship between the indices and DMW of soybeans. As a result, Σ (CI×I) was significantly related to the DMW, and had the least root square mean value (RMSE) of the three indices. The values of RMSE over the growth period were 98.5, 337.0, and 113.7 for Σ (CI×I) , RVI, and Σ (NDVI×I) , respectively. This was because the Σ (CI×I) in a reproductive period had the least RMSE of the three indices. These results indicate that the new vegetation index of Σ (CI×I) could estimate the dry matter weight of soybean more precisely than other indices over a soybean growth period.
View full abstract