TL;DR
This study investigates the complex relationship between land surface conditions (specifically, the evaporative fraction, EF) and afternoon precipitation across North America. Using a combination of observational and reanalysis data, the authors find that while some datasets suggest a strong link, precipitation persistence and atmospheric factors often confound this relationship, making it difficult to isolate the direct impact of land surface processes.
What Problem Does This Paper Address?
For decades, scientists have been interested in how soil moisture influences precipitation, particularly through its effect on the partitioning of energy at the land surface into latent and sensible heat fluxes, quantified by the evaporative fraction (EF). The hypothesis is that wetter soils lead to higher EF, which in turn could influence atmospheric stability and trigger convection, ultimately leading to more precipitation. This soil moisture-precipitation feedback is crucial for improving weather and seasonal forecasts. However, observational evidence for this feedback, especially the link between EF and precipitation, has been inconsistent and poorly understood. This paper aims to clarify this relationship by using multiple data sources and carefully accounting for confounding factors.
Methods
The core of the study involves calculating a metric called the "Triggering Feedback Strength" (TFS). This metric quantifies the statistical relationship between before-noon EF and the probability of afternoon precipitation. The authors employ a simplified version of TFS to allow for comparisons across different data sets. They also analyze the impact of precipitation persistence by examining the relationship between EF and precipitation on days with and without prior-day rainfall. Furthermore, they decompose the evaporative flux into its components (transpiration, bare soil evaporation, and interception evaporation) to investigate which specific land-surface processes might be driving the observed EF-precipitation relationships.
Datasets
The study utilizes a variety of datasets:
- North American Regional Reanalysis (NARR): A gridded reanalysis product providing atmospheric and land surface variables.
- FLUXNET: A global network of eddy-covariance sites providing direct measurements of surface turbulent heat fluxes (used to calculate EF).
- NEXRAD (NEXt generation weather RADar): A radar network providing precipitation data.
- GLEAM (Global Land Evaporation: the Amsterdam Methodology): A satellite-driven product estimating land surface evaporation and EF.
The analysis is primarily focused on North America, using data from 1995-2007.
Key Insights and Ideas
- Data Set Dependence is Significant: The observed EF-precipitation relationship varies considerably depending on the data source used for EF and precipitation. This highlights the substantial uncertainties in current EF estimates.
- Precipitation Persistence is a Major Confounder: The tendency for precipitation to persist from one day to the next often explains much of the apparent EF-precipitation relationship. This suggests that atmospheric conditions, rather than solely land surface processes, may be driving the observed correlations.
- Isolating Land's Role is Challenging: The study underscores the difficulty in disentangling the direct influence of land surface processes on precipitation from atmospheric influences and the feedback of precipitation on soil moisture.
- Regional Mechanisms Differ: The drivers of the EF-precipitation relationship vary across regions. In the eastern US, atmospheric controls and vegetation interception appear more dominant in NARR, while in the central and southwestern US, soil moisture plays a more significant role.
Key Results
- Eastern US Discrepancy: While NARR shows a strong positive EF-precipitation relationship in the eastern US, observational data (GLEAM-NEXRAD and FLUXNET-NEXRAD) do not confirm this. This suggests potential model artifacts in NARR or limitations in observational data.
- Central US Signal: The GLEAM-NEXRAD combination indicates a positive EF-precipitation relationship in the central US, which is consistent with expectations for a soil-moisture-limited regime.
- Southwestern US Agreement: All data sets generally agree on a positive EF-precipitation relationship in the southwestern US.
- Precipitation Predicts Better than EF: Previous day's precipitation is a stronger predictor of afternoon rainfall than before-noon EF, suggesting a short-timescale dominance of atmospheric processes.
- Interception's Role in NARR: In NARR, vegetation interception evaporation appears to play a significant role in the eastern US EF-precipitation relationship, though the authors suggest this might be overestimated due to model limitations.
Limitations and Future Work
- Data Uncertainties: Large uncertainties in EF data, particularly from satellite products, and limitations in FLUXNET's spatial scale and record length hinder definitive conclusions.
- Confounding Factors: While precipitation persistence is addressed, other confounding factors like large-scale atmospheric dynamics and land cover heterogeneity are not fully disentangled.
- Model Artifacts: The study points to potential model artifacts in NARR, particularly concerning interception evaporation, which may inflate observed relationships.
- Future Work: The authors emphasize the need for further studies with longer observational time series, a wider range of data sets, and analyses across different temporal and spatial scales. They also call for improved soil moisture and EF observations at scales relevant to land-atmosphere coupling to better constrain model results and advance our understanding of these complex interactions.
Summary updated