Using a network of networks for high-frequency multi-depth soil moisture observations to infer spatial and temporal drivers of subsurface preferential flow
Subsurface preferential flow (PF = water bypassing the soil matrix) provides rapid flowpaths for water and any substances transported with it, thereby profoundly impacting the recharge of aquifers, the spreading of contaminants, the health of the soil, and the functioning of ecosystems. It involves a complexity of processes that are poorly understood to the degree that current science provides no reliable way to predict its occurrence and magnitude. This effort will address the fundamental question of where and when PF occurs, taking advantage of two recent scientific developments: availability of high frequency (at least every 30 minutes), multi-depth soil moisture data suitable to detect preferential flow events across diverse meteorological conditions and a range of landscapes; and advances in artificial intelligence and machine learning (AI/ML). Outcomes from this synthesis will include: (1) a global database of harmonized high-frequency soil moisture and associated properties ready for AI applications, (2) AI/ML models that embody the control exerted on PF by weather and land characteristics, (3) a means to predict how the occurrence of PF will change with increased rainfall intensity and drought, and (4) approaches by which complex PF process can be represented in Earth System Models. The database and findings from this synthesis will enable further global analyses of various hydrological processes (e.g., streamflow) and serve as a benchmark data set for model development in Earth Science propelling the progress of hydrological modeling on various scales. This Powell Center Working Group is collaboratively supported by the NSF Critical Zone Network Hub and the U.S. Geological Survey.
Principal Investigators
Matthias Sprenger (Lawrence Berkeley National Laboratory)
Pamela Sullivan (Ohio State University)
John Nimmo (USGS Water Resources Mission Area)
Tianfan Xu (Arizona State University)
- Source: USGS Sciencebase (id: 6328ec6ed34e71c6d67b79e9)
Subsurface preferential flow (PF = water bypassing the soil matrix) provides rapid flowpaths for water and any substances transported with it, thereby profoundly impacting the recharge of aquifers, the spreading of contaminants, the health of the soil, and the functioning of ecosystems. It involves a complexity of processes that are poorly understood to the degree that current science provides no reliable way to predict its occurrence and magnitude. This effort will address the fundamental question of where and when PF occurs, taking advantage of two recent scientific developments: availability of high frequency (at least every 30 minutes), multi-depth soil moisture data suitable to detect preferential flow events across diverse meteorological conditions and a range of landscapes; and advances in artificial intelligence and machine learning (AI/ML). Outcomes from this synthesis will include: (1) a global database of harmonized high-frequency soil moisture and associated properties ready for AI applications, (2) AI/ML models that embody the control exerted on PF by weather and land characteristics, (3) a means to predict how the occurrence of PF will change with increased rainfall intensity and drought, and (4) approaches by which complex PF process can be represented in Earth System Models. The database and findings from this synthesis will enable further global analyses of various hydrological processes (e.g., streamflow) and serve as a benchmark data set for model development in Earth Science propelling the progress of hydrological modeling on various scales. This Powell Center Working Group is collaboratively supported by the NSF Critical Zone Network Hub and the U.S. Geological Survey.
Principal Investigators
Matthias Sprenger (Lawrence Berkeley National Laboratory)
Pamela Sullivan (Ohio State University)
John Nimmo (USGS Water Resources Mission Area)
Tianfan Xu (Arizona State University)
- Source: USGS Sciencebase (id: 6328ec6ed34e71c6d67b79e9)