Chin J Plant Ecol ›› 2006, Vol. 30 ›› Issue (4): 675-681.DOI: 10.17521/cjpe.2006.0088

• Research Articles • Previous Articles     Next Articles

ESTIMATION OF SOIL NITROGEN STATUS WITH CANOPY REFLECTANCE SPECTRA IN RICE

XUE Li-Hong, LU Ping, YANG Lin-Zhang*(), SHAN Yu-Hua, FAN Xiao-Hui, HAN Yong   

  1. Institute of Soil Science, Chinese Academy of Sciences, Nanjing 210008, China
  • Received:2005-01-12 Accepted:2005-10-30 Online:2006-01-12 Published:2006-07-30
  • Contact: YANG Lin-Zhang

Abstract:

Background and Aims Making nitrogen (N) recommendations without knowing the N supply capability of a soil can lead to inefficient use of N and potential pollution of the ground water. Conventional soil test techniques are destructive and time consuming. Remote sensing of canopy reflectance has the capacity of non-destructive and rapid estimating crop N status, and could be potentially used in evaluating soil N supply status.
Methods Thus, rice (Oryza sativa) canopy reflectance spectra and soil available N content (NH+4-N+NO-3-N) of different straw and N treatments were measured at key development stages of rice at two sites with different soil types and rice varieties. All possible normalized difference vegetation indices (NDVI) and ratio vegetation indices (RVI) composed by two bands, and some hybrid vegetation indices such as SAVI (soil adjusted vegetation index), TSAVI (transformed soil adjusted vegetation index) were calculated. Then the correlations between these VIs and soil available N were analyzed and the best regression equation was also investigated.
Key Results The correlations of soil available N content and canopy reflectance were negative at visible band, while positive at near infrared bands during the whole growing cycle. NDVI and RVI were well correlated with soil available N content, with the best stage of tillering and the best indices of the combination of 870, 1 220 nm and 560, 710 nm. Relationship between soil available N content and the best vegetation indices screened at tillering stage were influenced by soil background. While TSAVI was proved to be the best choice for removing the effect of soil background, and the relationship was consistent at two sites, with the best regression equation in exponential form. TSAVI calculated with the best band combinations screened in this study can improve the relationship, especially for the TSAVI calculated with 870 and 710 nm, with the decision coefficient (R2) increased from 0.46 to 0.60. At the heading and filling stage, the ratios of vegetation index and new SAVI calculated by 1 220 and 760 nm were well related to soil available N content independent of sites.
Conclusions Our observations suggest that evaluating soil N status at rice growing stage of with canopy reflectance spectra is feasible, but more research still need to be conducted to test and improve the soil N prediction model.

Key words: Canopy reflectance spectra, Soil available nitrogen content, Vegetation indices, Rice