Unverified paper record
Relating spatial turfgrass quality to actual evapotranspiration for precision golf course irrigation
Crop Science. · 1 Jan 2025 · 10.1002/csc2.21446
Abstract
Golf courses are increasingly affected by water scarcity and climate change. An understanding of spatial variability of actual evapotranspiration (ETₐ) and turfgrass quality (TQ) site-specific management zones (SSMZ) is important for the implementation of precision turfgrass management. Therefore, the main objectives of this study were to quantify the relationship between remotely sensed TQ and ETₐ estimates and to evaluate the spatial variations of TQ and ETₐ at a golf course in Utah. Ground-based normalized difference vegetation index was collected using a TCM-500 sensor, and aerial multispectral and thermal imagery data were acquired from unpiloted aircraft systems (UAS) in 2021, 2022, and 2023. A remote sensing TQ-random forest (RF) model was developed using six datasets of UAS spectral indices and the RF algorithm. The spatial data were analyzed to determine the correlation between TQ and ETₐ estimates. The TQ and ETₐ SSMZ were created and integrated with irrigation heads on the golf course using the Thiessen polygons tool. Results demonstrated that TQ-RF model was accurate within a root mean square error of 0.05. The correlation between TQ-RF and ETₐ was stronger for fairways (R² = 0.74), tees (R² = 0.66), and roughs (R² = 0.75) as compared to greens (R² = 0.25) and the driving range (R² = 0.36) on July 20, 2022. Actual evapotranspiration SSMZ, in combination with TQ-RF SSMZ, is useful for irrigation scheduling, addressing the question of how much and where to irrigate. This study demonstrates the ability of TQ-RF and ETₐ SSMZ to identify spatial variation for the purpose of landscape irrigation management in semi-arid areas.
Plant phenotyping relevance
UAS画像とRFモデルにより芝草品質という植物状態を推定する手法を開発・検証しており、表現型取得が研究の中心である。
abstractA remote sensing TQ-random forest (RF) model was developed using six datasets of UAS spectral indices and the RF algorithm.
abstractResults demonstrated that TQ-RF model was accurate within a root mean square error of 0.05.
Code and data availability
The supplied blocks contain only the article text, author contributions, and reference list. No public phenotype dataset, imagery, code repository, model checkpoint, or supplement URL with explicit availability language is provided; the Supporting Information note lacks any link or deposit identifier.
No evidence-backed public reproduction asset is currently recorded.
This is an automatically classified, unverified record. Curator approval is required before any resource enters the Catalog.