乌兰布和沙漠东北缘荒漠-绿洲过渡带植被地上生物量估算

干旱区荒漠植被地上生物量是植被生长状况评价与荒漠化监测的重要指标。在乌兰布和沙漠东北缘的荒漠-绿洲过渡带选取典型区,基于地面调查数据构建主要植物种的异速生长方程,对样方内的植被地上生物量进行估算;基于样方调查数据和Quick Bird影像数据,分别建立植被指数与人工固沙林和荒漠植被地上生物量的回归模型,并对研究区植被地上生物量进行估算。结果表明:植冠体积V是较好的预测变量,所得荒漠植物异速生长方程精度较高,能够满足样方内荒漠植被地上生物量估算需要;采用RVI对数模型估算人工固沙林地上生物量的效果最好(R~2=0.72,RMSEP=56.15),采用RVI线性模型估算荒漠植被地上生物量的效果最好(R~2=0.82,RMSEP=15.07);研究区内荒漠植被和人工固沙林的单位面积地上生物量分别为90.73g/m~2和105.28g/m~2。该研究可以为荒漠化监测和荒漠植被遥感信息提取提供参考。

Estimation of aboveground biomass of vegetation in the desert-oasis ecotone on the northeastern edge of the Ulan Buh Desert

Drylands in China are large in area and have an arid climate. The dominant vegetation is desert vegetation in drylands. Because desert vegetation is very sparse, it is difficult to extract information, such as vegetation cover and biomass, using remote sensing technology. Aboveground biomass of desert vegetation is an important indicator for evaluation of vegetation growth and desertification monitoring in dryland areas. A typical area on the desert-oasis ecotone was selected at the northeastern edge of the Ulan Buh Desert. Based on field investigation data, allometric equations of the primary plant species in the study area were established, and the aboveground biomass of vegetation in plots was estimated. Three vegetation indices, i.e., the ratio vegetation index (RVI), normalized difference vegetation index (NDVI), and modified soil adjusted vegetation index (MSAVI), were extracted using QuickBird image data as the remote sensing data source. Regression models of the vegetation indices and aboveground biomass of artificial sand-fixing woods and desert vegetation were established based on field investigation data, and then the aboveground biomass of vegetation in the study area was estimated. The results showed that:(1) the optimal predictive variable for the aboveground biomass allometric equation for the desert shrub species, such as Haloxylon ammodendron, Artemisia desertorum, Caragana korshinskii, Tamarix ramosissima, Nitraria tangutorum, and Elaeagnus angustifolia, was their crown volume, V, and all R² values for the equations were greater than 0.7, whereas for the desert shrub species Hedysarum scoparium the optimal predictive variable was its crown area, S, and the R² of the equation was 0.63; (2) among the vegetation index regression models constructed using RVI, NDVI, and MSAVI, the RVI logarithmic model was the best for the estimation of aboveground biomass of artificial sand-fixing woods (R² =0.72, RMSEP=56.15), whereas the RVI linear model was the best for estimating the aboveground biomass of desert vegetation (R² =0.82, RMSEP=15.07); (3) the aboveground biomass per unit area was 90.73g/m² and 105.28g/m² for desert vegetation and artificial sand-fixing woods, respectively. In the study area, the area of desert vegetation and artificial sand-fixing woods was 16.189km² and 15.685km², respectively, the total aboveground biomass was 3.12t and the aboveground biomass decreased from east to west. In this study, the allometric equation was used to estimate aboveground biomass of desert shrub species and high-resolution remote sensing data was adopted to assess the aboveground biomass of desert vegetation and artificial sand-fixing woods. The accuracy of the results was improved, especially for the sparse desert vegetation. This research will provide a reference for desertification monitoring and remote sensing information extraction of desert vegetation data. Further exploration of reliable methods for extracting information regarding sparse desert vegetation using multi-source remote sensing data is needed.