Fuzzy Correlation Evaluation of Forest EcologicalCarrying Capacity Considering the Interactionbetween Attributes

Jing Lu, Jian Lin, Ying Zhao, Yan HuangPolish 2023 · DOI

TL;DR

This paper introduces a sophisticated evaluation framework for Forest Ecological Carrying Capacity (FECC) that moves beyond simple additive models. By combining generalized trapezoidal fuzzy numbers with the $\lambda$-Shapley-Choquet integral, the researchers successfully account for the complex, non-linear interactions between resource indicators and environmental pressures across five Chinese provinces.

What Problem Does This Paper Address?

As economies rapidly expand, the pressure on forest ecosystems—our primary terrestrial carbon sinks and air purifiers—reaches critical thresholds. Measuring a region's "Forest Ecological Carrying Capacity" (FECC) is vital for sustainable development, yet traditional evaluation methods often fall short.

The core challenge is attribute interaction. Most existing models assume that indicators (like forest coverage and industrial pollution) are independent. In reality, these factors are deeply intertwined; for instance, a high forest stock can mitigate certain types of pollution, while severe pest outbreaks can drastically reduce a forest's ability to handle environmental stress. This paper seeks to bridge this gap by modeling these "fuzzy" interdependencies to provide a more accurate ranking of regional ecological health.

Methods

The authors propose a multi-stage mathematical pipeline to handle the uncertainty and complexity of ecological data:

  1. Generalized Trapezoidal Fuzzy Numbers (GTFNs): Instead of using simple averages for 17 years of data, the authors use GTFNs. This allows them to describe data with "flat peaks" (where values stay within a certain range for a long time) more accurately than standard triangular fuzzy numbers.
  2. $\lambda$-Fuzzy Measure & Generalized Shapley Value: Borrowing from cooperative game theory, the Shapley value is used to determine the "contribution" of each indicator. This helps assign weights not just to individual indicators, but to combinations of them, capturing how they interact.
  3. Optimal Linear Programming: To reduce subjectivity, the authors establish a linear programming model based on similarity to determine the fuzzy measures.
  4. $\lambda$-Shapley-Choquet Integral Operator: This is the "aggregator." It combines the fuzzy information into a comprehensive evaluation value while respecting the non-additive nature of the indicators.
  5. Centroid Method: Finally, the fuzzy results are converted into crisp values to rank the provinces.

Datasets

The study utilizes a 17-year panel dataset (2004–2020) focusing on five provinces in southeastern China: Zhejiang, Fujian, Jiangxi, Guangdong, and Guangxi.

  • Sources: Data was pulled from the 7th, 8th, and 9th National Forest Resource Inventories and the National Bureau of Statistics.
  • Indicators: 12 secondary indicators were used, categorized into:
  • Resource Indicators ($B_1$): Forest land area, coverage rate, stock volume, etc.
  • Damage Rate Indicators ($B_2$): Fire-affected areas and pest/rodent damage.
  • Environmental Pollution ($B_3$): COD discharge, ammonia nitrogen, sulfur dioxide, and industrial solid waste.
  • Preprocessing: Missing data points were forecasted using the ARIMA (Auto-Regressive Integrated Moving Average) method.

Key Insights and Ideas

  • Pollution is the Primary Bottleneck: Surprisingly, the generalized Shapley weight for the environmental pollution index ($B_3$) was the highest at 45%. This suggests that reducing industrial pressure is more critical for improving carrying capacity than simply planting more trees.
  • The "Fujian Advantage": Fujian Province consistently ranked first. This is attributed to its superior forest stock volume and high percentage of national nature reserves, which provide a robust buffer against environmental degradation.
  • Zhejiang’s "Damage" Hurdle: Despite strong economic resources, Zhejiang’s ranking was hampered by high scores in "damage rate" indicators (fire and pests), highlighting that ecological protection is as much about maintenance as it is about growth.
  • Preserving Information through Fuzziness: By using fuzzy numbers rather than crisp averages, the model retains the "history" and volatility of the 17-year period, leading to a more resilient evaluation.

Key Results

The comprehensive evaluation ranked the five provinces' forest ecological carrying capacity from highest to lowest as follows:

  1. Fujian Province (Centroid Value: 0.1853)
  2. Jiangxi Province (Centroid Value: 0.1597)
  3. Zhejiang Province (Centroid Value: 0.1456)
  4. Guangxi Province (Centroid Value: 0.1137)
  5. Guangdong Province (Centroid Value: 0.0998)

The results indicate that while Guangdong is an economic powerhouse, its high industrial output places a significant "load" on its forest ecosystems, resulting in the lowest carrying capacity among the group.

Limitations and Future Work

While the model is mathematically robust, the authors acknowledge that strengthening FECC requires a "parallel" approach: enriching forest resources while simultaneously reducing pollution.

Potential areas for future research include: * Granularity: Applying the model to city-level or county-level data for more localized policy-making. * Dynamic Interaction: Exploring how these interactions change over time (e.g., does the interaction between pests and pollution strengthen under climate change?). * Computational Complexity: The use of complex fuzzy domain roots and linear programming may require specialized software, which could limit its immediate adoption by non-technical forestry departments.

Summary updated

Abstract

Forest ecological carrying capacity can measure the balance between environmental pollution and forest ecosystems in the process of social development.This paper first selects 17-year panel data of five provinces in southeastern China, and uses the generalized trapezoidal fuzzy number to describe and express them based on the flat peak characteristic of the data itself; Secondly, the generalized Shapley value with the introduction of λ-fuzzy measure is used to measure the correlation between attribute indicators, and the indicator weight is determined by the attributes contribution; Then, an optimal fuzzy measure linear programming model is established based on the similarity, which is used to determine the generalized Shapley value of each attribute indicator.Finally, the λ-Shapley-Choquet integral operator is used for information aggregation and the centroid method is used to rank the comprehensive evaluation values.The results show that the forest ecological carrying capacity of five provinces during 2004-2020 is ranked from high to low as Fujian Province, Jiangxi Province, Zhejiang Province, Guangxi Province, and Guangdong Province.Based on this, it is proposed that strengthening the forest's ecological carrying capacity in southeastern China should be carried out in parallel from two aspects: enriching forest resources and reducing environmental pollution.
BibTeX
@inproceedings{lu2023fuzzy,
  author = {Jing Lu and Jian Lin and Ying Zhao and Yan Huang},
  title = {{Fuzzy Correlation Evaluation of Forest EcologicalCarrying Capacity Considering the Interactionbetween Attributes}},
  booktitle = {Polish},
  year = {2023},
  doi = {10.15244/pjoes/160195},
}
Figure