The grassland on semiarid loess hilly-gully region is an ecosystem with nutrient limitation.One of the main measures to promote the restoration of degraded grassland in this type of area is to increase soil nutrient content through exogenous N and P fertilization.Rational N and P combination can not only efficiently improve grassland productivity but also affect community species composition and diversity.Strengthening the monitoring of grassland growth after N and P addition is of great significance for evaluating the effect of exogenous fertilization.Aboveground biomass(AGB),cover,leaf area index(LAI),and diversity indexes are important parameters for vegetation growth status.Hyperspectral technology provides a new way for rapid monitoring of vegetation physiological and ecological parameters.This paper takes natural grassland communities in the loess hilly-gully region and their dominant species(L.davurica,A.mongolica,P.tanacetifolia,A.gmelinii,B.ischaemum,and S.bungeana)as research objects.By collecting and analyzing the spectral characteristics of grassland community and their dominant species under N and P addition in different months,combined with the correlation study between community characteristics and their spectral characteristics,sensitive spectral characteristics related to community characteristics were screened out and a model was established.The main research results are as follows:1.The spectral reflectance curves of the dominant species in different loess hilly-gully region and the FD curves within the red edge range are basically consistent,but there are obvious differences in the near-infrared band.In the near-infrared band,the spectral reflectance of A.mongolica is the largest(up to 0.71)and that of S.bungeana is the smallest among the six dominant species in the loess hilly-gully area.L.davurica,P.tanacetifolia,A.gmelinii,and B.ischaemum are in the middle.There are differences between the spectral characteristics of different dominant species,and the differences are more obvious after N and P addition.In the visible to near-infrared band features,FD and red edge parameters can be used to identify dominant species in loess hilly-gully region.2.The difference between the spectral reflectance of the canopy of the grass community under different N and P addition in the visible band is inversely proportional to the amount of fertilizer applied and in the near infrared band is directly proportional to the amount of fertilizer applied,the differences in the spectral reflectance of the community canopy became larger as the growth period progressed(May to August);the"double peak"phenomenon of the first-order derivative in the red-edge area is significantly affected by the N and P addition.The"double peak"phenomenon in the first-order derivative in the red-edge area is significantly influenced by N and P addition,and becomes more and more obvious as the amount of fertilizer applied increases and the growth period progresses.The differences in the spectral characteristics of grassland communities under different N and P addition become more and more obvious as plant growth and development progress.3.Among the 10 spectral feature value parameters composed of vegetation indices and red edge parameters,spectral feature parameters(NDVI,RVI,MASVI,SIPI,GNDVI,and AMP)that are more sensitive to chlorophyll and nutrient content play an important role in estimating community characteristics of grassland community during the vigorous growth period.The high spectral estimation model of Cover,LAI,and AGB(R_v~2 is 0.82,0.81,and 0.87,respectively)constructed is higher than the estimation accuracy of diversity indexes(Shannon-Wiener index,Simpson index,Patrick index,and Pielou index)(R_v~2 is 0.67,0.65,0.63,and0.61,respectively).This study uses hyperspectral technology to achieve rapid and accurate estimation of grassland community characteristics in the loess hilly-gully region,laying a foundation for large-scale monitoring of grassland community characteristics after N and P addition. |