The potential of vegetation indices: estimating methane emissions in paddy fields under nighttime warming and water saving conditions

Abstract Nondestructive and rapid methods for estimating methane emissions from paddy fields are helpful in indentifying options for emission mitigation, both at global and local scales. However, few studies have used an approach of vegetation indices for estimating methane emissions in paddy fields from ground‐based hyperspectral data. Therefore, this study systematically analyzed the quantitative relationship of methane emissions from rice fields to hyperspectral vegetation indices based on hyperspectral reflectance data of rice canopy. A field experiment was designed with two nighttime warming levels, that is, nighttime warming and control (ambient temperature), and two irrigation levels, that is, flooding (control, with 5 cm depth of water layer) and wetting (water saving irrigation, without water layer). Results showed that the newly developed three‐band index (R1139 − R960)/(R1139 + R960 − 2R944) (Rλ is the reflectance value at wavelength λ) was recommended as the optimal index for monitoring methane emission from paddy field, generated coefficients of determination, root mean square error, and residual prediction deviation values between the measured and predicted values of 0.687, 1.424 mg m−2 h−1, and 1.83, respectively. This study provides a theoretical basis and technical support for methane emissions assessment in paddy fields based on vegetation indices. © 2020 Society of Chemical Industry and John Wiley & Sons, Ltd.