Hyperspectral characteristics and growth monitoring of rice (Oryza sativa) under asymmetric warming

Author：

ISSN： 1366-5901

Source： International Journal of Remote Sensing, Vol.34, Iss.23, 2013-12, pp. : 8449-8462

Disclaimer: Any content in publications that violate the sovereignty, the constitution or regulations of the PRC is not accepted or approved by CNPIEC.

Previous Menu Next

Abstract

Climate warming shows great diurnal variation, with a higher warming rate at night-time, and consequently causes significant impacts on rice growth and development. Soil and plant analyser development (SPAD) values and leaf area index (LAI) are important parameters for indicating rice growth potential. The non-destructive, quick assessment of SPAD values and LAI is necessary for growth diagnosis in rice production. The objectives of this study were to determine the effects of asymmetric warming on rice growth and development, to define the relationships of SPAD values and LAI to ground-based canopy hyperspectral reflectance and derivative parameters, and to establish quantitative models for real-time monitoring of rice growth using sensitive spectral parameters under asymmetric warming. Two field warming experiments were performed in Nanjing in Jiangsu Province, China, to investigate the effects of asymmetric warming (all-day warming, AW; daytime warming from 7:00 to 19:00, DW; and night-time warming from 19:00 to 7:00, NW) on the growth and development of the Huaidao 5 rice cultivar, using a free air temperature increase (FATI) apparatus. The daily mean rice canopy temperatures in the AW, DW, and NW treatment plots were 2.0°C, 1.1°C, and 1.3°C higher, respectively, than those in the control (CK) plots. Canopy hyperspectral reflectance, SPAD, and LAI were measured under warming treatments during different growth stages. The results showed that rice SPAD and LAI decreased under warming treatments, with significant differences among certain growth stages compared with CK. The SPAD and LAI dynamics during growth stages exhibited single-peak patterns. Sensitive spectral bands occurred mostly within the visible light and near-infrared regions. An integrated linear regression equation relating SPAD to first-order derivative 514 (FD514) and ratio vegetation index (810,560) (RVI (810,560)) accurately described the dynamic pattern of SPAD changes, the coefficients of determination (R²) of 0.53 and 0.59, with the standard errors (SEs) of 0.85 and 1.13. The hyperspectral parameters first-order derivative 506 (FD506) and blue edge area (SDb) exhibited a linear relationship with LAI, with R² values of 0.79 and 0.80 and SE values of 0.69 and 0.23, respectively. When independent data were used to test the derived equations, the R² values between the measured and estimated SPAD from FD514 and RVI (810,560) were 0.72 and 0.60, with root mean square errors (RMSEs) of 4.34 and 4.44 and relative errors (REs) of 11.00% and 12.00%, respectively. For spectral parameters FD506 and SDb, the R² values between the measured and estimated LAI values were 0.75 and 0.75, with RMSEs of 0.57 and 0.48 and REs of 17.00% and 4.00%, respectively. The strong fit between the measured and estimated values indicated that the present models based on hyperspectral reflectance could be used for reliable SPAD and LAI estimation under asymmetric warming.