Productive restoration planning in the irrigated drylands: Soil salinity, texture and moisture to guide reforestation plans
- 13 February 2026
- journal article
- Published by Wiley in Journal of Applied Ecology
- Vol. 63 (2)
Abstract
Soil salinity is one of the most widespread causes of declines in agricultural yields in drylands, consequently threatening environmental sustainability and food security. Productive restoration is an opportunity to repair the structure and function of these agroecosystems. Among possible restoration measures, afforestation with native trees ( Neltuma spp., ‘algarrobos’) is a promising alternative, particularly in the Cuyo oasis (Argentina). Although the relevance of restoration planning has been widely recognised, there is a knowledge gap regarding the planning of productive restoration in agroecosystems. Our objective is to identify potential sites for reforestation in the irrigated drylands of San Juan. To that end, we used biomass productivity prediction based on soil salinity, texture, and moisture derived from a reforestation experiment. We evaluated bands from individual Sentinel‐2 acquisitions to account for these soil properties. We selected them based on their correlation with field sample data to use as predictor variables. Biomass productivity was modelled using Random Forests, with the growing‐season integral of the Normalised Difference Vegetation Index red edge (iNDVIre) as a proxy for productivity and 29 predictor variables. Based on variable importance, RMSE, and MAE, we selected a subset of seven predictors. We performed a spatial 3‐fold cross‐validation for the final model, and R 2 was 0.68. We identified 134 potential restoration sites affected by soil salinity and capable of supporting the highest forest biomass values. Sites totalled 248.50 ha, representing 10% of the area most affected by soil salinity in this sector of the oasis. Sites varied in size and were distributed across the entire study area. Biomass productivity prediction can be used to plan restoration actions in agroecosystems affected by soil salinity to ensure a higher probability of success and lower costs. Moreover, the proposed approach enabled the scaling up of experimental findings. Synthesis and applications . The proposed approach is a valuable tool for productive restoration planning in irrigated drylands and is adaptable to other agroecosystems constrained by different soil factors.Keywords
This publication has 34 references indexed in Scilit:
- Supervised machine learning for predicting and interpreting dynamic drivers of plantation forest productivity in northern Tasmania, AustraliaComputers and Electronics in Agriculture, 2023
- Rehabilitation planning in an oilfield in the Monte Austral: Mapping sand‐sized sediment availability and assessing its effect on microtopography rehabilitationEarth Surface Processes and Landforms, 2022
- Critical knowledge gaps and research priorities in global soil salinity2021
- Ecological Restoration and Productive Recovery of Saline Environments from the Argentine Monte Desert Using Native Plants2020
- Satellite data integration for soil clay content modelling at a national scaleInternational Journal of Applied Earth Observation and Geoinformation, 2019
- Soil Salinity: Historical Perspectives and a World Overview of the Problem2018
- Planning restoration in human-modified landscapes: New insights linking different scalesApplied Geography, 2017
- Priority areas for forest restoration aiming at the conservation of water resourcesEcological Engineering, 2016
- Predicting site index of plantation loblolly pine from biophysical variablesForest Ecology and Management, 2014
- Detecting salinity hazards within a semiarid context by means of combining soil and remote-sensing dataGeoderma, 2006