Land subsidence hazard modeling

Machine learning to identify predictors and the role of human activities

Omid Rahmati, Ali Golkarian*, Trent Biggs, Saskia Keesstra, Farnoush Mohammadi, Ioannis N. Daliakopoulos

*Corresponding author for this work

Research output: Contribution to journalArticleAcademicpeer-review

10 Citations (Scopus)

Abstract

Land subsidence caused by land use change and overexploitation of groundwater is an example of mismanagement of natural resources, yet subsidence remains difficult to predict. In this study, the relationship between land subsidence features and geo-environmental factors is investigated by comparing two machine learning algorithms (MLA): maximum entropy (MaxEnt) and genetic algorithm rule-set production (GARP) algorithms in the Kashmar Region, Iran. Land subsidence features (N = 79) were mapped using field surveys. Land use, lithology, the distance from traditional groundwater abstraction systems (Qanats), from afforestation projects, from neighboring faults, and the drawdown of groundwater level (DGL) (1991–2016) were used as predictive variables. Jackknife resampling showed that DGL, distance from afforestation projects, and distance from Qanat systems are major factors influencing land subsidence, with geology and faults being less important. The GARP algorithm outperformed the MaxEnt algorithm for all performance metrics. The performance of both models, as measured by the area under the receiver-operator characteristic curve (AUROC), decreased from 88.9–94.4% to 82.5–90.3% when DGL was excluded as a predictor, though the performance of GARP was still good to excellent even without DGL. MLAs produced maps of subsidence risk with acceptable accuracy, both with and without data on groundwater drawdown, suggesting that MLAs can usefully inform efforts to manage subsidence in data-scarce regions, though the highest accuracy requires data on changes in groundwater level.

Original languageEnglish
Pages (from-to)466-480
JournalJournal of Environmental Management
Volume236
DOIs
Publication statusPublished - 15 Apr 2019

Fingerprint

Subsidence
Learning systems
Groundwater
Hazards
subsidence
human activity
hazard
drawdown
groundwater
modeling
genetic algorithm
afforestation
Genetic algorithms
entropy
Land use
Entropy
groundwater abstraction
Lithology
land
machine learning

Keywords

  • Groundwater overexploitation
  • Iran
  • Land use change
  • Subsidence
  • Sustainability

Cite this

Rahmati, Omid ; Golkarian, Ali ; Biggs, Trent ; Keesstra, Saskia ; Mohammadi, Farnoush ; Daliakopoulos, Ioannis N. / Land subsidence hazard modeling : Machine learning to identify predictors and the role of human activities. In: Journal of Environmental Management. 2019 ; Vol. 236. pp. 466-480.
@article{9b7471ebe74f4d5ab62aec962aacc4bd,
title = "Land subsidence hazard modeling: Machine learning to identify predictors and the role of human activities",
abstract = "Land subsidence caused by land use change and overexploitation of groundwater is an example of mismanagement of natural resources, yet subsidence remains difficult to predict. In this study, the relationship between land subsidence features and geo-environmental factors is investigated by comparing two machine learning algorithms (MLA): maximum entropy (MaxEnt) and genetic algorithm rule-set production (GARP) algorithms in the Kashmar Region, Iran. Land subsidence features (N = 79) were mapped using field surveys. Land use, lithology, the distance from traditional groundwater abstraction systems (Qanats), from afforestation projects, from neighboring faults, and the drawdown of groundwater level (DGL) (1991–2016) were used as predictive variables. Jackknife resampling showed that DGL, distance from afforestation projects, and distance from Qanat systems are major factors influencing land subsidence, with geology and faults being less important. The GARP algorithm outperformed the MaxEnt algorithm for all performance metrics. The performance of both models, as measured by the area under the receiver-operator characteristic curve (AUROC), decreased from 88.9–94.4{\%} to 82.5–90.3{\%} when DGL was excluded as a predictor, though the performance of GARP was still good to excellent even without DGL. MLAs produced maps of subsidence risk with acceptable accuracy, both with and without data on groundwater drawdown, suggesting that MLAs can usefully inform efforts to manage subsidence in data-scarce regions, though the highest accuracy requires data on changes in groundwater level.",
keywords = "Groundwater overexploitation, Iran, Land use change, Subsidence, Sustainability",
author = "Omid Rahmati and Ali Golkarian and Trent Biggs and Saskia Keesstra and Farnoush Mohammadi and Daliakopoulos, {Ioannis N.}",
year = "2019",
month = "4",
day = "15",
doi = "10.1016/j.jenvman.2019.02.020",
language = "English",
volume = "236",
pages = "466--480",
journal = "Journal of Environmental Management",
issn = "0301-4797",
publisher = "Elsevier",

}

Land subsidence hazard modeling : Machine learning to identify predictors and the role of human activities. / Rahmati, Omid; Golkarian, Ali; Biggs, Trent; Keesstra, Saskia; Mohammadi, Farnoush; Daliakopoulos, Ioannis N.

In: Journal of Environmental Management, Vol. 236, 15.04.2019, p. 466-480.

Research output: Contribution to journalArticleAcademicpeer-review

TY - JOUR

T1 - Land subsidence hazard modeling

T2 - Machine learning to identify predictors and the role of human activities

AU - Rahmati, Omid

AU - Golkarian, Ali

AU - Biggs, Trent

AU - Keesstra, Saskia

AU - Mohammadi, Farnoush

AU - Daliakopoulos, Ioannis N.

PY - 2019/4/15

Y1 - 2019/4/15

N2 - Land subsidence caused by land use change and overexploitation of groundwater is an example of mismanagement of natural resources, yet subsidence remains difficult to predict. In this study, the relationship between land subsidence features and geo-environmental factors is investigated by comparing two machine learning algorithms (MLA): maximum entropy (MaxEnt) and genetic algorithm rule-set production (GARP) algorithms in the Kashmar Region, Iran. Land subsidence features (N = 79) were mapped using field surveys. Land use, lithology, the distance from traditional groundwater abstraction systems (Qanats), from afforestation projects, from neighboring faults, and the drawdown of groundwater level (DGL) (1991–2016) were used as predictive variables. Jackknife resampling showed that DGL, distance from afforestation projects, and distance from Qanat systems are major factors influencing land subsidence, with geology and faults being less important. The GARP algorithm outperformed the MaxEnt algorithm for all performance metrics. The performance of both models, as measured by the area under the receiver-operator characteristic curve (AUROC), decreased from 88.9–94.4% to 82.5–90.3% when DGL was excluded as a predictor, though the performance of GARP was still good to excellent even without DGL. MLAs produced maps of subsidence risk with acceptable accuracy, both with and without data on groundwater drawdown, suggesting that MLAs can usefully inform efforts to manage subsidence in data-scarce regions, though the highest accuracy requires data on changes in groundwater level.

AB - Land subsidence caused by land use change and overexploitation of groundwater is an example of mismanagement of natural resources, yet subsidence remains difficult to predict. In this study, the relationship between land subsidence features and geo-environmental factors is investigated by comparing two machine learning algorithms (MLA): maximum entropy (MaxEnt) and genetic algorithm rule-set production (GARP) algorithms in the Kashmar Region, Iran. Land subsidence features (N = 79) were mapped using field surveys. Land use, lithology, the distance from traditional groundwater abstraction systems (Qanats), from afforestation projects, from neighboring faults, and the drawdown of groundwater level (DGL) (1991–2016) were used as predictive variables. Jackknife resampling showed that DGL, distance from afforestation projects, and distance from Qanat systems are major factors influencing land subsidence, with geology and faults being less important. The GARP algorithm outperformed the MaxEnt algorithm for all performance metrics. The performance of both models, as measured by the area under the receiver-operator characteristic curve (AUROC), decreased from 88.9–94.4% to 82.5–90.3% when DGL was excluded as a predictor, though the performance of GARP was still good to excellent even without DGL. MLAs produced maps of subsidence risk with acceptable accuracy, both with and without data on groundwater drawdown, suggesting that MLAs can usefully inform efforts to manage subsidence in data-scarce regions, though the highest accuracy requires data on changes in groundwater level.

KW - Groundwater overexploitation

KW - Iran

KW - Land use change

KW - Subsidence

KW - Sustainability

U2 - 10.1016/j.jenvman.2019.02.020

DO - 10.1016/j.jenvman.2019.02.020

M3 - Article

VL - 236

SP - 466

EP - 480

JO - Journal of Environmental Management

JF - Journal of Environmental Management

SN - 0301-4797

ER -