Soil Memory and Dryland Restoration Failure: Legacy Effects, Functional Thresholds, and Recovery Pathways: Review Article

Restoration projects in drylands fail in more than one way: some never establish, some decline after early success, and some gain vegetation cover while soil and hydrological functions remain impaired. Across 174 dryland seeding sites worldwide, 17% established none of the species sown, and seeded-species occurrence declined as projects aged. Thisreview examines whether persistent soil legacies can help explain part of this variation. Soil memory is used here in arestricted sense: a soil-resident imprint of past degradation, disturbance, management, or climatic stress that can belinked to degradation history, persists beyond short-term or seasonal variability after the original pressure has declined,and can modify the response to restoration. Evidence from soil science, dryland ecology, ecohydrology, restoration ecology, and rangeland research is synthesized against this definition. The review uses a purposive, mechanism-based synthesis rather than a systematic review or meta-analysis, and SMRF remains a framework for prospective testing rather than a validated decision model. Six domains are distinguished: chemical, physical, biological, hydrological, management-imprinted, and coupled memory. Current grazing, tenure, access, enforcement, and maintenance are treated separately as feasibility conditions. Five failure patterns are identified, ranging from immediate establishment failure to implementation failure. The Soil Memory-Based Restoration Framework (SMRF) is proposed as a diagnosis-led sequencethat links soil-memory classification, threshold assessment, feasibility screening, and restoration pathway selection.Remote sensing supports screening and stratification, while soil-memory classification requires field or laboratoryconfirmation. Where a historical reference state is no longer feasible under current and projected climate, SMRFseparates recovery of soil function from endpoint feasibility. Indicators and published numerical values are used as locally interpreted screening evidence rather than universal thresholds. The framework may therefore help restoration planning distinguish visible recovery from recovery of the soil, hydrological, biological, and vegetation functions needed for persistence.