Agroforestry Features of Restoration of Degraded Lands in the Conditions of the Guangdong Province

Abstract

This dissertation addresses the critical issue of land degradation in Guangdong Province, China's most economically dynamic region, which faces a unique contradiction between high-intensity development and inherent ecological fragility. The research establishes a comprehensive "macro - micro - technology - model" framework to develop scientifically optimized and regionally adapted agroforestry solutions for restoring degraded red soils. The theoretical foundation analyzes the global severity of land degradation and its specific manifestations in China and Guangdong, identifying the region's distinctive degradation mechanisms: physical erosion from high temperature and heavy rainfall averaging 1777 mm annually, chemical pollution from industrialization, and biological degradation from urbanization. Agroforestry is established as a nature-based solution for simultaneously restoring soil health, enhancing biodiversity, and providing economic benefits. The methodology integrates multi-scale approaches including macro-level remote sensing using MODIS and Landsat data with GIS spatial analysis and the RUSLE model to assess spatiotemporal degradation patterns from 2000 to 2020; technology optimization utilizing a hybrid Random Forest - Response Surface Methodology (RF - RSM) model to quantify synergistic effects of biochar, lime, and leguminous green manure; and micro - mechanism investigation applying high - throughput sequencing and FAPROTAX functional prediction to analyze soil microbial community responses. A comprehensive statistical inference protocol was implemented, including Shapiro - Wilk normality testing, Levene's test for variance homogeneity, ANOVA with Duncan's multiple range post - hoc comparisons (p < 0.05), and PERMANOVA with 999 permutations for multivariate analyses. The research reveals accelerating land degradation in the Pearl River Delta, with highly degraded areas surging by 196.58% from 2000 to 2020 and particularly rapid deterioration after 2010 showing a 97.83% increase. Soil erosion dominates the natural degradation process with 40% weight, spatially coupled with the land degradation index at R2 = 0.91. Human factors shifted from early industrial expansion between 2000 and 2010, during which 82.6% of flat farmland was covered by factories, to later ecological policy regulation, demonstrating the effectiveness of the 2014 Ecological Control Line Plan which reduced severe desertification by 4% in protected zones. The developed and validated RF - RSM hybrid model for optimizing agronomic measures identified through random forest analysis that available phosphorus with importance score 2.10 and soil pH with importance score 1.85 serve as core driving factors. Response surface methodology quantified optimal combinations of biochar at 8 to 12 tons per hectare, lime at 3 tons per hectare, and milkvetch green manure. The optimized scheme reduced available cadmium from 0.42 mg/kg to 0.11 mg/kg, representing a 73.8% reduction, increased soil porosity to 65.3%, and raised pH from 5.0 to 7.5. Cost - effectiveness improved dramatically as cadmium passivation cost decreased from 35.93 USD per percentage point using traditional methods to 9.19 USD per percentage point, with rebound rate controlled at only 3.8%. Elucidation of microscopic mechanisms through soil microbiome analysis using high - throughput sequencing revealed that green manure application increased bacterial Chao1 richness by 23.6 - 60% and Shannon diversity by 7 - 28.1% under controlled pot conditions. Proteobacteria and Acidobacteria became dominant phyla, with ryegrass increasing Firmicutes abundance threefold. FAPROTAX functional prediction suggested significantly enhanced inferred nitrogen fixation, nitrification, fermentation, and chemoheterotrophic functional potentials. These inferred functional shifts require direct metatranscriptomic or enzymatic validation to confirm actual metabolic activity. Redundancy analysis identified pH, total nitrogen, and soil organic matter as core environmental drivers shaping microbial communities. These findings provide mechanistically grounded hypotheses for field restoration; however, their quantitative applicability to open - field agroecosystems requires dedicated field validation in future research. Meta - analysis of 183 effect sizes from 15 published studies confirmed green manure increases crop yield by 13.8% and soil organic carbon by 17.3%. The synthesis of findings resulted in the design of three regionally differentiated agroforestry models as scientifically grounded pilot frameworks. The Pearl River Delta Rapid Improvement Model for heavy metal pollution achieves remediation within one year with 78.9% cadmium stabilization. The Eastern and Western Guangdong Forest - Fruit - Grass Composite Model for soil erosion reduces erosion modulus by 70 - 90%. The Northern Guangdong Near - Natural Transformation Model for ecological barrier zones increases water conservation capacity by 20 – 40 %. These proposed models showed synergistic production, ecology, and economy benefits in pilot trials, with land equivalent ratio increasing 30 - 50% and carbon sequestration potential 20 - 35% above monocultures. In conclusion, this research provides a full - chain scientific framework from degradation pattern recognition to precision restoration technology optimization, establishing replicable agroforestry solutions for subtropical degraded red soils. The integrated approach achieves optimal balance between ecological restoration effectiveness and economic feasibility, offering theoretical guidance and practical support for sustainable land management in Guangdong Province and similar regions globally. Importantly, all experimental agronomic findings are empirically grounded in an 18 - month field monitoring program (supplemented by controlled pot experiments). Long - term projections - including 3 - year reapplication cycles and carbon stability beyond 5 years - are presented as model - based hypotheses requiring future verification and should be interpreted with appropriate caution.