Using machine learning to produce near surface soil moisture estimates from deeper in situ records at U.S. Climate Reference Network (USCRN) locations: Analysis and applications to AMSR-E satellite validation

Evan J. Coopersmith, Michael H. Cosh, Jesse E. Bell, Ryan Boyles

Research output: Contribution to journalArticle

2 Citations (Scopus)

Abstract

Surface soil moisture is a critical parameter for understanding the energy flux at the land atmosphere boundary. Weather modeling, climate prediction, and remote sensing validation are some of the applications for surface soil moisture information. The most common in situ measurement for these purposes are sensors that are installed at depths of approximately 5 cm. There are however, sensor technologies and network designs that do not provide an estimate at this depth. If soil moisture estimates at deeper depths could be extrapolated to the near surface, in situ networks providing estimates at other depths would see their values enhanced. Soil moisture sensors from the U.S. Climate Reference Network (USCRN) were used to generate models of 5 cm soil moisture, with 10 cm soil moisture measurements and antecedent precipitation as inputs, via machine learning techniques. Validation was conducted with the available, in situ, 5 cm resources. It was shown that a 5 cm estimate, which was extrapolated from a 10 cm sensor and antecedent local precipitation, produced a root-mean-squared-error (RMSE) of 0.0215 m3/m3. Next, these machine-learning-generated 5 cm estimates were also compared to AMSR-E estimates at these locations. These results were then compared with the performance of the actual in situ readings against the AMSR-E data. The machine learning estimates at 5 cm produced an RMSE of approximately 0.03 m3/m3 when an optimized gain and offset were applied. This is necessary considering the performance of AMSR-E in locations characterized by high vegetation water contents, which are present across North Carolina. Lastly, the application of this extrapolation technique is applied to the ECONet in North Carolina, which provides a 10 cm depth measurement as its shallowest soil moisture estimate. A raw RMSE of 0.028 m3/m3 was achieved, and with a linear gain and offset applied at each ECONet site, an RMSE of 0.013 m3/m3 was possible.

Original languageEnglish (US)
Pages (from-to)122-131
Number of pages10
JournalAdvances in Water Resources
Volume98
DOIs
StatePublished - Dec 1 2016

Fingerprint

AMSR-E
soil moisture
climate
sensor
network design
climate prediction
shallow soil
analysis
machine learning
in situ
energy flux
in situ measurement
water content
remote sensing
weather
atmosphere
vegetation
resource

Keywords

  • AMSR-E
  • ECONet
  • Modeling
  • Remote sensing
  • Soil moisture
  • USCRN

ASJC Scopus subject areas

  • Water Science and Technology

Cite this

@article{514af5a49ca5452e866afdf71a4b0fa8,
title = "Using machine learning to produce near surface soil moisture estimates from deeper in situ records at U.S. Climate Reference Network (USCRN) locations: Analysis and applications to AMSR-E satellite validation",
abstract = "Surface soil moisture is a critical parameter for understanding the energy flux at the land atmosphere boundary. Weather modeling, climate prediction, and remote sensing validation are some of the applications for surface soil moisture information. The most common in situ measurement for these purposes are sensors that are installed at depths of approximately 5 cm. There are however, sensor technologies and network designs that do not provide an estimate at this depth. If soil moisture estimates at deeper depths could be extrapolated to the near surface, in situ networks providing estimates at other depths would see their values enhanced. Soil moisture sensors from the U.S. Climate Reference Network (USCRN) were used to generate models of 5 cm soil moisture, with 10 cm soil moisture measurements and antecedent precipitation as inputs, via machine learning techniques. Validation was conducted with the available, in situ, 5 cm resources. It was shown that a 5 cm estimate, which was extrapolated from a 10 cm sensor and antecedent local precipitation, produced a root-mean-squared-error (RMSE) of 0.0215 m3/m3. Next, these machine-learning-generated 5 cm estimates were also compared to AMSR-E estimates at these locations. These results were then compared with the performance of the actual in situ readings against the AMSR-E data. The machine learning estimates at 5 cm produced an RMSE of approximately 0.03 m3/m3 when an optimized gain and offset were applied. This is necessary considering the performance of AMSR-E in locations characterized by high vegetation water contents, which are present across North Carolina. Lastly, the application of this extrapolation technique is applied to the ECONet in North Carolina, which provides a 10 cm depth measurement as its shallowest soil moisture estimate. A raw RMSE of 0.028 m3/m3 was achieved, and with a linear gain and offset applied at each ECONet site, an RMSE of 0.013 m3/m3 was possible.",
keywords = "AMSR-E, ECONet, Modeling, Remote sensing, Soil moisture, USCRN",
author = "Coopersmith, {Evan J.} and Cosh, {Michael H.} and Bell, {Jesse E.} and Ryan Boyles",
year = "2016",
month = "12",
day = "1",
doi = "10.1016/j.advwatres.2016.10.007",
language = "English (US)",
volume = "98",
pages = "122--131",
journal = "Advances in Water Resources",
issn = "0309-1708",
publisher = "Elsevier Limited",

}

TY - JOUR

T1 - Using machine learning to produce near surface soil moisture estimates from deeper in situ records at U.S. Climate Reference Network (USCRN) locations

T2 - Analysis and applications to AMSR-E satellite validation

AU - Coopersmith, Evan J.

AU - Cosh, Michael H.

AU - Bell, Jesse E.

AU - Boyles, Ryan

PY - 2016/12/1

Y1 - 2016/12/1

N2 - Surface soil moisture is a critical parameter for understanding the energy flux at the land atmosphere boundary. Weather modeling, climate prediction, and remote sensing validation are some of the applications for surface soil moisture information. The most common in situ measurement for these purposes are sensors that are installed at depths of approximately 5 cm. There are however, sensor technologies and network designs that do not provide an estimate at this depth. If soil moisture estimates at deeper depths could be extrapolated to the near surface, in situ networks providing estimates at other depths would see their values enhanced. Soil moisture sensors from the U.S. Climate Reference Network (USCRN) were used to generate models of 5 cm soil moisture, with 10 cm soil moisture measurements and antecedent precipitation as inputs, via machine learning techniques. Validation was conducted with the available, in situ, 5 cm resources. It was shown that a 5 cm estimate, which was extrapolated from a 10 cm sensor and antecedent local precipitation, produced a root-mean-squared-error (RMSE) of 0.0215 m3/m3. Next, these machine-learning-generated 5 cm estimates were also compared to AMSR-E estimates at these locations. These results were then compared with the performance of the actual in situ readings against the AMSR-E data. The machine learning estimates at 5 cm produced an RMSE of approximately 0.03 m3/m3 when an optimized gain and offset were applied. This is necessary considering the performance of AMSR-E in locations characterized by high vegetation water contents, which are present across North Carolina. Lastly, the application of this extrapolation technique is applied to the ECONet in North Carolina, which provides a 10 cm depth measurement as its shallowest soil moisture estimate. A raw RMSE of 0.028 m3/m3 was achieved, and with a linear gain and offset applied at each ECONet site, an RMSE of 0.013 m3/m3 was possible.

AB - Surface soil moisture is a critical parameter for understanding the energy flux at the land atmosphere boundary. Weather modeling, climate prediction, and remote sensing validation are some of the applications for surface soil moisture information. The most common in situ measurement for these purposes are sensors that are installed at depths of approximately 5 cm. There are however, sensor technologies and network designs that do not provide an estimate at this depth. If soil moisture estimates at deeper depths could be extrapolated to the near surface, in situ networks providing estimates at other depths would see their values enhanced. Soil moisture sensors from the U.S. Climate Reference Network (USCRN) were used to generate models of 5 cm soil moisture, with 10 cm soil moisture measurements and antecedent precipitation as inputs, via machine learning techniques. Validation was conducted with the available, in situ, 5 cm resources. It was shown that a 5 cm estimate, which was extrapolated from a 10 cm sensor and antecedent local precipitation, produced a root-mean-squared-error (RMSE) of 0.0215 m3/m3. Next, these machine-learning-generated 5 cm estimates were also compared to AMSR-E estimates at these locations. These results were then compared with the performance of the actual in situ readings against the AMSR-E data. The machine learning estimates at 5 cm produced an RMSE of approximately 0.03 m3/m3 when an optimized gain and offset were applied. This is necessary considering the performance of AMSR-E in locations characterized by high vegetation water contents, which are present across North Carolina. Lastly, the application of this extrapolation technique is applied to the ECONet in North Carolina, which provides a 10 cm depth measurement as its shallowest soil moisture estimate. A raw RMSE of 0.028 m3/m3 was achieved, and with a linear gain and offset applied at each ECONet site, an RMSE of 0.013 m3/m3 was possible.

KW - AMSR-E

KW - ECONet

KW - Modeling

KW - Remote sensing

KW - Soil moisture

KW - USCRN

UR - http://www.scopus.com/inward/record.url?scp=84993945343&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84993945343&partnerID=8YFLogxK

U2 - 10.1016/j.advwatres.2016.10.007

DO - 10.1016/j.advwatres.2016.10.007

M3 - Article

AN - SCOPUS:84993945343

VL - 98

SP - 122

EP - 131

JO - Advances in Water Resources

JF - Advances in Water Resources

SN - 0309-1708

ER -