From Reactive Remediation to Predictive Prevention: A Systems Synthesis of Global Soil Salinization and Sodicity

Authors:
Open Access
Publisher website
Google Scholar
Abstract
Soil salinization and sodicity are accelerating worldwide, reducing crop productivity across more than 160 million hectares of farmland and causing production losses approaching USD 27 billion annually. Despite decades of research documenting these impacts, the extent of salt‐affected soils continues to expand. Salt‐affected soil degradation emerges from feedbacks among physical, biological, and socioeconomic processes that operate across scales and resist intervention when addressed in isolation. This review synthesizes current understanding of these coupled feedbacks. We trace how ion accumulation, soil structural decline, microbial community disruption, and plant stress interact with irrigation practices, land use decisions, and governance systems. These interactions create threshold dynamics in which degradation can appear stable for years before rapid decline. Evidence from diverse regions reveals that effective restoration depends on coordinated interventions matched to the complexity of the system itself. Four research priorities emerge for advancing prediction and prevention. First, quantifying the thresholds that separate reversible from recalcitrant degradation across soil types and climates. Second, expanding predictive capacity through integrated sensor networks and data platforms that enable early warning. Third, improving understanding of how institutional structures influence adoption of restoration practices. Fourth, integrating biophysical and governance models to enable scaling from field trials to basin‐wide implementation. By coupling mechanistic understanding with monitoring infrastructure and adaptive governance, reactive remediation can give way to proactive prevention.
Funding Information
  • National Institute of Food and Agriculture (IOW04108)
  • Novo Nordisk Fonden (NNF19SA0059360)

This publication has 73 references indexed in Scilit: