China’s Forest Ecosystem Service Value Climbed 12 Percent Over Two Decades, Study Finds

The monetary value of China’s forest ecosystem services rose by 12 percent between 2000 and 2023, according to a new study that tracked the nation’s forests county by county across a continuous annual series. The research, published in Frontiers in Forests and Global Change, applied the equivalent factor method with coefficient adjustments at both provincial and county levels, allowing the authors to follow how the estimated value of services such as climate regulation, water retention, and timber production shifted across the country over nearly a quarter century. Regulating services dominated the valuation in every single year, accounting for more than 66 percent of the national total, a figure that underscores how much of the estimated economic worth of Chinese forests lies not in harvestable products but in the quieter work of stabilizing climate and water systems.

The team, led by Qi Shi of Fujian Business University and Chuan Ye of Fujian Agriculture and Forestry University, relied on a battery of multisource datasets to build their valuation. Land-cover information came from the China Land Cover Dataset developed at Wuhan University at 30-meter resolution, while vegetation greenness was captured through an annual normalized difference vegetation index, or NDVI, dataset at 1-kilometer resolution provided by the Chinese Academy of Sciences. Grain production and average grain prices from national statistical yearbooks supplied the monetary baseline for the equivalent factor method, which uses the value of cropland food production as a reference unit for pricing ecosystem services. The researchers included both forest and shrubland in their definition of the forest ecosystem, and excluded Hong Kong, Macao, and Taiwan from the analysis.

What distinguishes the study methodologically is its two-tier coefficient adjustment. National baseline equivalent factors, averaged across four forest subtypes from a widely used reference table, were first scaled at the provincial level using the ratio of provincial to national grain yield per unit area, a proxy for regional cropland productivity. Then, at the county level, standardized NDVI indices were compared between each county and its province, and that ratio further adjusted the coefficients. This second step is intended to capture within-province differences in forest condition driven by tree species, stand age, and site conditions. The monetary value of one standard equivalent factor was set at 1,640.45 CNY per hectare, averaged across multiple periods to dampen the influence of annual price swings. County-level estimates were then aggregated upward to produce provincial and national figures for each year.

The temporal trajectory was far from linear. Total forest ecosystem service value hit its lowest point in 2001, climbed with fluctuations to a peak in 2018, and ended the period 12 percent above its starting level. The underlying forest area followed a similar wobble: it increased 1.51 percent between 2003 and 2011, dipped 0.46 percent from 2011 to 2014, then recovered by 1.16 percent through 2023, when it reached its highest level of the study period and covered 25.24 percent of China’s land area, up from 24.70 percent in 2000. Within the valuation framework, climate regulation and hydrological regulation carried the highest estimated values among individual services, while provisioning value reflected timber and non-timber forest products and cultural services, the smallest share, captured the aesthetic worth of forest landscapes.

Geographically, the results split in a revealing way. Western China held the largest total value, contributing more than 46 percent of the national figure on average and increasing its share from 45.93 percent in 2000 to 49.06 percent in 2023, a gain of 19.64 percent in absolute terms. Northeastern China’s total rose 19.31 percent over the same span. By contrast, eastern China’s total declined 2.06 percent and central China’s fell 0.62 percent. Yet when the authors measured value per unit of forest area, the ranking flipped: eastern China ranked first throughout the study period, exceeding western China by between 26.12 and 53.05 percent annually, with an average gap of 39.13 percent, though that difference narrowed from 50.25 percent in 2000 to 29.98 percent in 2023. A region can hold vast forests of moderate productivity, the findings show, while another holds smaller forests of strikingly higher per-hectare value.

Classifying provinces into five value tiers using fixed thresholds derived from 2018 data made the regional structure vivid. Inner Mongolia, Heilongjiang, Sichuan, and Yunnan occupied the highest tier from 2013 onward, and the number of top-tier provinces grew from two in 2000 to four in 2023, all clustered in western and northeastern China where forest extent is largest. At the bottom, Beijing, Tianjin, Shanghai, Jiangsu, Shandong, Qinghai, Ningxia, and Xinjiang remained in the lowest class in every selected year, a pattern the authors associate with limited forest ecosystem area rather than any single natural or land-use factor. Between these poles, provinces such as Jilin and Shaanxi moved up a tier during the early 2000s and held their new positions, while Liaoning, Hubei, Tibet, and Chongqing drifted between adjacent classes.

Spatial statistics added another layer to the picture. A standard deviational ellipse fitted to the national distribution showed a stable northeast–southwest orientation throughout the period, becoming slightly more elongated over time, which the authors interpret as a marginally stronger directional pattern. The center of gravity of the value distribution stayed within Henan Province, moving from near the Henan–Hubei boundary toward central Henan, a net displacement of 88.42 kilometers between 2000 and 2023, predominantly northward with a slight eastward component. The year-by-year trajectory was jagged: a 35.31-kilometer northeastward shift from 2000 to 2003, a further 66.66 kilometers northeast from 2003 to 2008, a northwestward turn through 2013, a 52.09-kilometer southwestward move to 2018, and another northwestward drift in the final five years.

To probe what drives the spatial variation, the authors turned to Geodetector, a statistical framework that compares variance across an entire study area with variance within strata defined by candidate explanatory factors. Using a 0.5-degree grid that yielded 3,846 sampling points, they tested eight factors in four representative years: precipitation, temperature, humidity, slope, and elevation on the natural side; population density, nighttime light intensity, and per capita GDP on the socioeconomic side. All eight showed statistically significant explanatory power in every year examined, at the 5 percent level. Topography reigned supreme among natural factors, with slope producing q-values between 0.3675 and 0.3963 and elevation between 0.3040 and 0.3200, while among socioeconomic factors, per capita GDP and population density carried relatively strong explanatory power, with their relative rankings shifting across years.

Perhaps the most striking analytical result concerned interactions. Every pairwise combination of factors exhibited either bivariate enhancement or nonlinear enhancement, meaning that stratifying the data jointly by two factors always explained more of the spatial variation than either factor alone. Combinations pairing elevation or slope with other natural variables produced particularly high q-values. The authors caution, however, that Geodetector quantifies explanatory power without identifying the direction of associations or confirming causal mechanisms, and that interaction detection measures changes in explanatory power after joint stratification rather than verifying a synergistic process. They also note that management variables such as stand age were omitted for lack of continuous observations, and that the analysis should be read as an initial indication of relative factor importance rather than a causal accounting.

The authors link the upward national trend to China’s major ecological programs, including natural forest protection, cropland-to-forest conversion, and the Three-North Shelterbelt Forest Program, alongside the nationwide collective forest tenure reform launched in 2008, which a 2024 National Forestry and Grassland Administration report credited with expanding collectively owned forest area by 37 percent relative to pre-reform levels. They suggest the northward drift of the value center of gravity may reflect afforestation and vegetation recovery in northern China, while southern provinces such as Fujian and Zhejiang, already rich in forest cover, have limited room for expansion and depend more on improvements in stand quality. The authors also flag the method’s limits: coefficients averaged across forest types, a single NDVI-based adjustment applied uniformly to all services, and a grain-based monetary baseline that would rescale all estimates uniformly if changed. The estimates, they stress, are best suited to relative comparisons across regions and years rather than absolute readings of ecological worth. Even so, the county-level annual series offers land managers a practical map of where forest value is concentrated, where it is thinning, and where conservation attention might be aimed next.

Subject of Research: Spatiotemporal dynamics and drivers of forest ecosystem service value in China from 2000 to 2023 using the equivalent factor method

Article Title: Spatiotemporal dynamics and drivers of forest ecosystem service value in China based on the equivalent factor method

Article References: Shi, Q., & Ye, C. (2026). Spatiotemporal dynamics and drivers of forest ecosystem service value in China based on the equivalent factor method. Frontiers in Forests and Global Change, 9, Article 1956763. https://doi.org/10.3389/ffgc.2026.1956763

Image Credits: AI Generated

DOI: 10.3389/ffgc.2026.1956763

Keywords: forest ecosystem services, China, equivalent factor method, ecosystem service valuation, spatiotemporal analysis, Geodetector, center-of-gravity model, regulating services, NDVI, forest conservation, land cover, spatial heterogeneity