Changes of NPP and their relationship to climate factors based on the transformation of different scales in Gansu, China

• The southeast of annual NPP is high, but low in the northwest, of Gansu.
• Annual NPP of forest and grassland biomes were influenced mainly by temperature.
• Annual NPP of desert and agricultural biomes were influenced mainly by precipitation.
• The monthly NPP showed a time-lag and cumulative effect in grassland biome.
• The monthly NPP showed a time-lag and cumulative effect in agricultural biome.

Net primary production (NPP) is one of the important indexes of ecological change. In this study, the spatial distribution of NPP in Gansu province was simulated using an improved Carnegie-Ames-Stanford Approach (CASA) model, from 2000 to 2010, using MODIS-NDVI data at 250 × 250 m spatial resolution. We analyzed NPP responses in forest, grassland, agricultural and desert biomes to climate change, by combining the simulated data with meteorological data. The results indicated that: (i) The spatial distribution of the annual NPP from 2000 to 2010 varied significantly over the region, with NPP high in the southeast, but low in the northwest, of Gansu. The mean annual NPP values for forest, grassland, agricultural and desert biomes were 580.12 gC m− 2 yr− 1, 380.12 gC m− 2 yr− 1, 280.71 gC m− 2 yr− 1 and 83.27 gC m− 2 yr− 1, respectively. (ii) The annual NPP values with extremely significant increase (ESI, θslope > 0 and p < 0.01) and significant increase (SI, θslope > 0 and 0.01 ≤ p < 0.05) from 2000 to 2010 accounted for 6.00% and 17.49% of the total area, respectively. The annual NPP values with extremely significant decrease (ESD, θslope < 0 and p < 0.01) and significant decrease (SD, θslope < 0 and 0.01 ≤ p < 0.05) from 2000 to 2010 accounted for 12.31% and 27.07% of the total area, respectively. Overall, the NPP values displayed a decreasing trend. (iii) The forest and grassland biomes were influenced mainly by temperature, while precipitation was the main controlling factor in the desert and agricultural biomes, at the annual scale. (iv) Compared with the forest and desert biomes, the grassland and agricultural biomes showed time-lag and cumulative correlations with precipitation and temperature, respectively.