Contrasting spatial variations between above and below-ground net primary productivity in global grasslands
Grassland net primary productivity (NPP) plays a crucial role in global terrestrial carbon cycle and carbon balance. However, the spatial pattern and influencing factors of NPP and its components, especially the belowground net primary productivity (BNPP) remain unclear. This study compiled a datase...
Saved in:
| Main Authors: | , , , , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Elsevier
2025-01-01
|
| Series: | Ecological Indicators |
| Subjects: | |
| Online Access: | http://www.sciencedirect.com/science/article/pii/S1470160X25000123 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | Grassland net primary productivity (NPP) plays a crucial role in global terrestrial carbon cycle and carbon balance. However, the spatial pattern and influencing factors of NPP and its components, especially the belowground net primary productivity (BNPP) remain unclear. This study compiled a dataset of 4194 field site measurements of NPP, ANPP and BNPP from 1940 to 2019, as well as climate, soil, geographic, and vegetation index data, to quantify the global patterns and potential mechanisms of the NPP and its components. Machine learning methods were used to identify the determinants of grassland productivity variables and to estimate global productivity at a 0.05° resolution. The results indicated that NPP and ANPP increased significantly along the precipitation gradient under various temperature conditions, particularly between 0 °C and 10 °C. While the BNPP/ANPP showed an opposite trend, highlighting a shift in biomass allocation from belowground to aboveground along the climate gradient. The relationship between NPP and temperature exhibited varying trends under different precipitation conditions. In drier regions (MAP < 500 mm), a threshold effect was observed, while in more humid regions (MAP > 500 mm), the positive effect was more pronounced. The interaction of climate, soil and vegetation indices had the most significant explanatory power for productivity. Among the nine machine learning methods compared, the random forest model provided the highest accuracy for productivity estimation (R2 = 0.63 ∼ 0.89). ANPP is higher than BNPP in tropical regions, while BNPP is higher in temperate and cold regions. Our findings underscore the spatial heterogeneity effect of temperature and precipitation in shaping global patterns of NPP. The high spatial resolution products of NPP and its components produced in our study may serve as valuable benchmarks for land models in simulating NPP and its allocation in response to climate change. |
|---|---|
| ISSN: | 1470-160X |