Global Patterns of Vegetation Response to Short-Term Surface Water Availability

Qing He,Hui Lu,Kun Yang,Ling Zhen,Siyu Yue,Yishan Li,Dara Entekhabi

DOI: https://doi.org/10.1109/jstars.2021.3103854

IF: 4.715

2021-01-01

IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing

Abstract:Terrestrial vegetation response to surface water availability is important for land–atmosphere interactions. However, the current understanding of how the vegetation responds to surface water remains limited since the physical processes happening within the biosphere and hydrosphere are highly coupled. It is even more difficult to measure such interactions for the processes related to surface soil moisture (SSM)—the central variable that interacts the most intimately with vegetation—since the observations of SSM are often scarce and uneven. Here, we use the satellite observations of vegetation optical depth (VOD) and SSM to map the response time scales of vegetation to surface water anomalies. We use the stability theory to derive vegetation memory time ( ${{{\bf \tau }}_{{{\bf ReS}}}}$ ) to reveal the global pattern of vegetation memory to surface water anomalies. That is, the time vegetation takes to return back to its equilibrium when an anomaly dissipates to a certain level (e.g., the e-folding level). We also estimate the plant reactive time ( ${{{\bf \tau }}_{{{\bf ReA}}}}$ )—the time when impacts of surface anomaly reach its peak to evaluate the overall resilience of terrestrial vegetation to surface water anomalies. The results show that ${{{\bf \tau }}_{{{\bf ReS}}}}$ tends to be longer in herbaceous biomes, whereas ${{{\bf \tau }}_{{{\bf ReA}}}}$ is longer in biomes with tree cover. Such anticorrelation of ${{{\bf \tau }}_{{{\bf ReS}}}}$ and ${{{\bf \tau }}_{{{\bf ReA}}}}$ indicates that the herbaceous biomes may be more vulnerable to surface water perturbations during climate extremes. Our study provides a global quantification on vegetation—soil moisture feedbacks—enabling comparison with earth system models.