Initial Response of Normalized Difference Vegetation Index, Green Normalized Difference Vegetation Index and Soil Adjusted Vegetation Index to Infrared Warming in Highland Barley of the Tibet
Climatic warming affects the crop growth and its related vegetation indices. In order to understand the initial response of normalized difference vegetation index (NDVI), normalized green difference vegetation index (GNDVI) and soil adjusted vegetation index (SAVI) to climatic warming, a field warming experiment using infrared radiator was conducted in a highland barley located at the Dazi county of the Tibet since late May, 2015. There were three warming treatments, i.e., control, low (1000 W) and high(2000W) warming. The NDVI, GNDVI and SAVI were obtained using an agricultural digital camera during the period from June to September in 2015. Meanwhile, the soil temperature and soil moisture at depths of 5 cm and 20 cm were also obtained using HOBO microclimate observing systems. Then this study analyzed the response of NDVI, GNDVI and SAVI to infrared warming and the relationships between the three vegetation indices and soil temperature and moisture. The 1000 W and 2000 W infrared warming increased soil temperature at the depth of 5 cm (t_5) by 1.62 ℃ and 1.77 ℃, and soil temperature at the depth of 20 cm (t_(20)) by 1.16 ℃ and 1.43 ℃, but decreased soil moisture at the depth of 5 cm (SM5) by 1.8% and 14.1%, and soil moisture at the depth of 20 cm (SM_(20)) by 21.6% and 14.7%, respectively. The 1000 W infrared warming increased NDVI by 2.4%, GNDVI by 4.3% and SAVI by 0.5%, whereas the 2000 W infrared warming increased NDVI by 5.5%, GNDVI by 5.3% and SAVI by 4.8%, although these changes were non-significant. Simple regression analyses showed that (1) NDVI (r~2=0.110, P=0.026) and GNDVI(r~2=0.254, P=0.000 4)decreased with increasing t_5, whereas there was non-significant correlation between SAVI and t_5 (r~2=0.069, P=0.082); (2) GNDVI decreased with increasing t_(20), (r~2=0.218, P=0.001), whereas there were non-significant relationships between NDVI(r~2=0.040, P=0.190), SAVI (r~2=0.014, P=0.443) and t_(20); (3) NDVI (r~2=0.277, P=0.000 2), GNDVI (r~2=0.394, P=0.000 0) and SAVI(r~2=0.208, P=0.002)increased with increasing SM5; and (4) GNDVI increased with increasing SM20 (r~2=0.193, P=0.003), whereas there were non-significant correlations between NDVI (r~2=0.059, P=0.107), SAVI (r~2=0.037, P=0.209) and SM_(20). Multiple regression analyses indicated that SM5 dominated the variations of NDVI, GNDVI and SAVI. Partial correlation analyses demonstrated that (1) the correlation coefficients of NDVI, GNDVI and SAVI with SM5 were 0.442 (P=0.003), 0.412 (P=0.007) and 0.404 (P=0.008); (2) with SM_(20)were -0.042 (P=0.792), 0.051 (P=0.749) and -0.033 (P=0.837); (3) with t_5 were -0.154 (P=0.332), -0.019 (P=0.907) and -0.170 (P=0.282); and (4) with t_(20) were 0.228 (P=0.147), -0.041 (P=0.795) and 0.268 (P=0.086), respectively. Therefore, the soil drying induced by infrared warming suppressed the growth of highland barley, which in turn affected vegetation indices. That is, the non-significant changes of the three vegetation indices may be due to the infrared warming-induced drying.