MétaCan
Menu
Back to cohort
Record W4411665539 · doi:10.1111/ejss.70142

Applying <scp>LSTM</scp> to Model Multi‐Depth Soil Moisture Under Various Land Covers, Climates and Soils

2025· article· en· W4411665539 on OpenAlex
Nasrin Azad, Amirreza Sheikhbaglou, Francis Zvomuya, Hailong He

Why this work is in the frame

A frame that forgets how it found something cannot be audited. These are the routes that admitted this work.

affAt least one author lists a Canadian institution in the pinned OpenAlex snapshot.
fundA Canadian funder is recorded on the work.

Bibliographic record

VenueEuropean Journal of Soil Science · 2025
Typearticle
Languageen
FieldEnvironmental Science
TopicSoil Moisture and Remote Sensing
Canadian institutionsUniversity of Manitoba
FundersHigh-end Foreign Experts Recruitment Plan of ChinaNatural Sciences and Engineering Research Council of CanadaMinistry of Human Resources and Social SecurityNational Natural Science Foundation of China
KeywordsSoil waterEnvironmental scienceMoistureWater contentSoil scienceHydrology (agriculture)GeologyGeotechnical engineeringMeteorologyGeography

Abstract

fetched live from OpenAlex

ABSTRACT Accurate estimation of multi‐depth/profile soil moisture (SM) is required for sustainable water management in agriculture and hydrology. However, monitoring SM is costly and labour‐intensive, and only limited soil depths can be instrumented with soil moisture sensors. Therefore, various numerical simulation and data assimilation techniques have been used in multi‐depth soil moisture estimation. Machine learning (ML) has also gained popularity in SM estimation due to its ease of use and robustness, although proper handling of ML models also requires expertise and experience. However, the applicability of ML to estimate time series of multi‐depth SM under different land uses is mainly limited by the choice of ML models and the availability of SM data. In addition, the reliability of the trained model remains unknown when it is applied to different locations. Therefore, the objective of this study was to evaluate the widely used Long Short‐Term Memory (LSTM) model to estimate multi‐depth SM under different land covers, climates, and soils. A minimum of 10 years' daily meteorological and soil data at multiple depths were collected from six U.S. Climate Reference Network (USCRN) stations with different land covers and various climates and soils. These data were used to train the LSTM model and optimize its input parameters. Performance of the trained LSTM model was evaluated for multi‐depth SM estimation at two other “monitoring” stations with similar conditions. SM modeling at shallow depths (e.g., 5, 10 and 20 cm) was most accurate (&lt; 10% mean absolute percent error, MAPE) with precipitation and antecedent time series of SM as inputs, while the best SM estimates at deeper depths (e.g., 50 and 100 cm) were attained with antecedent SM time series as the input. Generation of the trained LSTM model from one station to other stations emphasized on the similar soil and land cover conditions. It is hoped that this research would provide better understandings of multi‐depth SM modeling and offer new insights improving profile SM modeling accuracy for un‐instrumented sites.

Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.

Full frame distilled prediction

Teacher imitation

Not calibrated prevalence, not ground truth. Human validation pending. Learned from the 10,348 direct Codex labels and 10,348 direct Gemma labels. Candidate is the union of thresholded teacher heads; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels or direct frontier model labels.

metaresearch head score (Codex)0.002
metaresearch head score (Gemma)0.000
Version: codex-gemma-dda1882f352aValidation status: machine_predicted_unvalidated
Candidate categoriesnone
Consensus categoriesnone
DomainCandidate signal: none · Consensus signal: none
Study designCandidate signal: Observational · Consensus signal: none
GenreCandidate signal: Empirical · Consensus signal: Empirical
Teacher disagreement score0.802
Threshold uncertainty score0.698

Codex and Gemma teacher scores by category

CategoryCodexGemma
Metaresearch0.0020.000
Meta-epidemiology (narrow)0.0000.000
Meta-epidemiology (broad)0.0000.000
Bibliometrics0.0000.001
Science and technology studies0.0010.001
Scholarly communication0.0000.000
Open science0.0010.001
Research integrity0.0000.000
Insufficient payload (model declined to judge)0.0000.000

Machine scores (provisional)

The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.

Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.

Opus teacher head0.013
GPT teacher head0.239
Teacher spread0.225 · how far apart the two teachers sit on this one work
Validation statusscore_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from it