Economic-mathematical approaches to the valuation of ecosystem services in the context of sustainable regional development

Abstract

The study is devoted to the development and verification of an economic-mathematical approach to the monetary valuation of ecosystem services as a foundation for sustainable regional development. Traditional macroindicators ignore the contribution of natural capital and exacerbate the uncertainty of non-market valuations. The aim is to create a hybrid algorithm and an integrated information-analytics system, EcoValue Modeller, to obtain comparable and scenario-sensitive valuations. The novelty lies in combining market and non-market methods within a single spatiotemporal database, followed by sensitivity and scenario analyses. The methods rely on integrating data for 2015-2024: regional statistics, satellite remote sensing data from Landsat 8 and Sentinel-2, hydrological data, and survey results, supplemented by NLP analysis of media. Implementation is in Python using Pandas, GeoPandas, scikit-learn, and TensorFlow. The algorithm combines replacement cost for regulating services; a modified travel cost method with hedonic pricing for recreation; and contingent valuation with logistic regression for biodiversity. We performed cross-validation, sensitivity analysis with respect to the discount rate and the CO2 price, comparisons with avoided-cost and travel-cost methods, and scenario modeling through 2035 along three trajectories. The results show that aggregate value is higher in the forested region due to the dominance of carbon sequestration, whereas in the urbanized region recreation and water purification prevail. Uncertainty is highest for biodiversity in the former and for recreation in the latter. Sensitivity to the discount rate and CO2 price is higher in the forested region. The hybrid model yields more comprehensive estimates than narrow single-method approaches and reduces the risk of underestimation. Scenario analysis reveals degradation under a business-as-usual path, loss of natural capital under an industry-driven growth path, and increases in value and GRP under a sustainable strategy. The discussion confirms the need for differentiated natural-capital management policies and a shift to dynamic, risk-oriented planning. The suite enhances the applicability of valuations thanks to method hybridization and spatiotemporal integration. Limitations relate to data quality, parametric uncertainty, and the need for regular model recalibration. Next steps include expanding observation sources, implementing MLOps, real-time monitoring dashboards, and a deeper accounting for sociocultural preferences. The approach operationalizes natural capital and supports the selection of effective portfolios of measures.