Dynamic change of net primary productivity during process of new artificial oasis expansion in Xinjiang based on remote sensing

Abstract: The impact of human activities on the global carbon cycle is one of the main concerns in the global climate change research. However, the mechanisms of the carbon cycle for the different regions, natural and human factors are very complex and different. Exploring the impact of human reclamation activities on the distribution pattern of net primary productivity (NPP) in arid oasis ecosystem will help to increase the understanding of human activities on terrestrial ecosystems and carbon cycle impacts in arid regions. This study aimed to investigate the change of NPP during the process of new artificial oasis expansion in Xinjiang based on the remote sensing data. June was selected as the optimal time for extracting vegetation cover types change information. Based on the long-term Landsat remote sensing data, the Carnegie-Ames-Stanford Approach (CASA) model was used to estimate the change of NPP during the expansion of Xinjiang Qitai artificial oasis from June 1976 to June 2016. The results showed that: 1) In 1976, the study area mainly was dominated by desert steppe, and the farmland was only 5 150 hm2, accounting for 3.25% of the area. By 2016, the amount of farmland expanded to 64 675hm2, which increased by more than 11 times; 2) While the artificial oasis was expanding, spring overflow zone disappeared, the swampland and the reservoir dried up, which resulted in the decrease of shrub area from 17 871 hm2 to 5 202 hm2; 3) Reclamation resulted in a significant decline in groundwater level, soil salinity development process was interrupted, and bare alkaline area was reduced from 58 859 hm2 in 1976 to 17 803 hm2 in 2016, which was reduced by nearly 70% and began to succession to the new grassland; 4) In June 1976, the unit area NPP of farmland was 68.6 g/m2, and the unit area NPP of farmland in June 2007 and June 2016 was over 90 g/m2. In the recent 40 years, the unit area NPP of grassland was fluctuated within the range of 30-40 g/m2, which was relatively stable; 5) In June 1976, the total NPP in the study area was 41 018 t, of which the contribution rate of farmland was only 8.62%. In June 2016, the total NPP of the study area reached to 90 602 t, which was 2.2 times higher than that in June 1976. The contribution rate of farmland to total NPP was 65.23%. In the same period, the net loss of shrub and swamp NPP was 8 265 t and 3 319 t, respectively; 6) Human activities were the decisive factors that influenced the distribution pattern of NPP in the study area. And NPP in 1976 were mainly composed of natural grassland and shrub, accounting for 56.68% and 26.61% respectively, NPP in 2016 was mainly composed of farmland and fresh grassland, accounting for 65.23% and 31.85%, respectively. In 1976, 66.93% of the shrubs, 75.58% of the swamps, 41.43% of the grassland, 23.86 % of the bare alkali land and 77.96% of the water bodies were evolved into farmland in 2016. The 43.62% of the grassland in 1976 was still grassland after 40 years, but was divided into fragments by the newly reclaimed farmland. Only 15.01% of the bare alkaline area in 1976 remained in 2016, while the rest came from degradation of grasslands, shrubs and swamps. In the recent 40 years, the original natural oasis landscape in the study area had been completely changed, and large areas of farmlands connected to each other become modern agriculture oasis that was suitable for large-scale agricultural machinery operations.