The township-scale analysis for agricultural residues treatment and resource recovery in rural China

China’s rural regions generate large volumes of heterogeneous organic residues, including crop straw, livestock manure, and smaller fractions of green and kitchen waste, yet their treatment and utilization remain highly uneven across townships. Anaerobic digestion (AD) and composting are the most mature pathways for stabilizing organic matter, recovering nutrients, and producing renewable energy; however, evaluating these technologies at township scale is constrained by coarse official statistics, spatially dispersed feedstocks, and limited local operational expertise. Effective planning therefore requires transparent, data-lean tools that integrate process modelling with spatial and system-level assessments. This dissertation develops and applies such a framework by addressing three research questions: (RQ1) How should composting/ AD models be selected to simulate the fate of carbon, nitrogen, and phosphorus to balance accuracy, interpretability, and cross-township transferability under township-level data unavailability; (RQ2) what is the spatially explicit, realistic potential for biogas-to-electricity from local crop residues; and (RQ3) what is the realistic potential of composting and AD for township-level sustainable agricultural development, considering nutrient cycling? A multi-stage methodology was established. First, a systematic review of 22 composting models and 52 AD models evaluates their nutrient-tracking capabilities, calibration requirements, and applicability under sparse data conditions. Mechanistic models offer high interpretability for C/N/P processes but demand extensive calibration, while machine-learning approaches achieve predictive accuracy only when supported by rich site-specific datasets. Second, three methane-yield modelling strategies, Biochemical Methane Potential (BMP), simplified ADM1 (S-ADM1), and machine-learning regression, were calibrated using field and laboratory data from Beiyang Township (Zhejiang Province). These models were integrated into a 50×50 m GIS framework to produce China’s first high-resolution township-scale biogas maps, identifying optimal digester siting locations and translating residue availability into electricity-generation potential. Third, a township-level Material Flow Analysis compares baseline, AD, and composting scenarios. Results show low baseline nutrient recovery (C: 8%, N: 3.6%, P: 1.1%) and high losses (N: 67.6%, P: 77.3%). Optimized AD scenarios improve C and N recovery to 18.4% and 16.9%, respectively, while composting yields the highest P recovery (5.5%) and lowest C loss (7.4%). Integrated strategies reduce dependence on synthetic fertilizers, enhance soil fertility, and generate 777.5 Mg of biogas and 849.3 Mg of composted carbon annually. Overall, the dissertation demonstrates that the township represents a crucial but understudied scale for agricultural residue management. By linking process models with spatial analysis and system-level nutrient accounting, it offers a transferable, data-efficient decision framework that supports rural energy transitions, circular-economy objectives, and sustainable agricultural development in China.