Modeling aboveground net primary production using Landsat-8 indices in Siahpoosh and Ganjgah rangelands of Ardabil province, Iran

Subject Areas : Agriculture, rangeland, watershed and forestry

Pashmineh  Mohammadnazhad ¹ (Department of Range and Watershed Management, Faculty of Agriculture and Natural Resources, University of Mohaghegh Ardabili, Ardabil, Iran)
Mehdi  Moameri ² (Department of Plant Science and Medicinal Plants, Faculty of MeshkinShahr Agriculture, University of Mohaghegh Ardabili, Ardabil, Iran)
Ardavan  Ghorbani ³ (Department of Range and Watershed Management, Faculty of Agriculture and Natural Resources, University of Mohaghegh Ardabili, Ardabil,)
Farid  Dadjou ⁴ (Department of Range and Watershed Management, Faculty of Agriculture and Natural Resources, University of Mohaghegh Ardabili, Ardabil, Iran)
Vadood  Mohammadi ⁵ (Department of Range and Watershed Management, Faculty of Agriculture and Natural Resources, University of Mohaghegh Ardabili, Ardabil, Iran)

Received: 2022-02-27 Accepted : 2022-03-29 Published : 2023-05-22

Keywords: Ardabil rangelands, Vegetation indices, LANDSAT 8, Life forms,

Abstract :

The aim of this study was to estimate the aboveground net primary production (ANPP) of life forms and total ANPP using Landsat 8 images in Siahpoosh and Ganjgah rangelands of Ardabil province. Field sampling was conducted in June 2019 and a satellite image was received at the same time. Seven sampling sites was selected, and three 100-meter transects parallel and perpendicular to the slope direction was established in each sites, and along each transects from 10 plots (1m2), the ANPP of life forms with random-systematic method were collected (210 plots in total). Then, 22 plant indices were selected according to previous references and calculated for the region. Next, the correlation between the ANPP of life forms and total ANPP with plant indices was calculated and the plant index with the highest correlation was selected for modeling. The first-order linear equation was used for modeling and the equations were estimated in ArcMap software as a map. The results showed that among the studied indices, NDVI was the most appropriate index for modeling; however, the highest correlation of this index was with total ANPP (0.88), and grasses ANPP (0.78); while it had less correlation with the forbs ANPP (0.41), and shrubs ANPP (0.31). The range of changes were 0 to 1857 kg/ha for grasses, 9 to 766 kg/ha for forbs, 0 to 458 kg/ha for shrubs, and 9 to 3081 kg/ha for the total ANPP. The accuracy of the models was evaluated by RMSE, MDE and MAE criteria and the accuracy was acceptable. Also, there was essentially no difference between the mean of real and modeled data. The results of this study can be used to balance the supply and demand of rangeland production for sustainable development of rangeland ecosystems.