Importance of the antecedent environmental factors’ memory effects on the temporal variation of terrestrial gross primary productivity
Quantitative estimation of temporal variation in ecosystem productivity is crucial for assessing the stability and sustainability of ecosystem carbon sinks. However, current assessments of temporal variation of gross primary productivity (GPP) suffer from inaccuracies due to oversight of the memory...
Saved in:
| Main Authors: | , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Elsevier
2025-06-01
|
| Series: | Ecological Indicators |
| Subjects: | |
| Online Access: | http://www.sciencedirect.com/science/article/pii/S1470160X25004881 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | Quantitative estimation of temporal variation in ecosystem productivity is crucial for assessing the stability and sustainability of ecosystem carbon sinks. However, current assessments of temporal variation of gross primary productivity (GPP) suffer from inaccuracies due to oversight of the memory effect of GPP on antecedent environmental and vegetation changes. By introducing memory effect into a time-dependent deep learning model, we investigated the responses of GPP to antecedent environmental and vegetation factors, and further simulated and analyzed the temporal trend and interannual variation of GPP at site and spatial scales. Our results indicate that (i) incorporating memory effect significantly improves the explanatory power of environmental and vegetation factors on GPP magnitude, trend, and interannual variation compared to the model ignoring memory effect; (ii) the memory effect length of GPP response to antecedent environmental and vegetation factors varies across different ecosystems, ranging from 4 to 11 months. Precipitation has a longer cumulative effect on GPP compared to temperature, shortwave radiation and VPD (Vapor Pressure Deficit) in most ecosystems. The impact of NDVI (Normalized Difference Vegetation Index) on GPP was stronger than environmental variables, emphasizing the significance of vegetation state in GPP simulation; (iii) the global terrestrial ecosystem GPP estimated by the deep learning model considering memory effect showed an increasing trend and significant interannual variation from 1983 to 2015. This study enhanced the understanding on the driving mechanisms of antecedent environmental and vegetation factors on GPP and provided a reference for modeling of carbon cycle process. |
|---|---|
| ISSN: | 1470-160X |