[Objective] To quantify key ecosystem services within the Dongting Lake region, clarify the trade-off/synergy relationships, and detect the driving factors in order to support the ecological sustainable development of the Dongting Lake region. [Methods] Using the InVEST model, taking the area around Dongting Lake as the study area, four ecosystem services including water yield, carbon storage, soil conservation, and habitat quality were quantitatively assessed. Interdependencies between ecosystem services were assessed using correlation analysis to quantify trade-offs/synergies, and the geodetector model was used to detect their driving factors. [Results] (1) From 2000 to 2020, the soil retention service and water yield service in the Dongting Lake area showed an increasing trend over time. The total water yield increased from 4.93 × 10¹⁰ m³ to 6.71 × 10¹⁰ m³, while the total soil retention increased from 4.46 × 10⁹ t to 5.77 × 10⁹ t; habitat quality and total carbon storage continued to decline, with habitat quality decreasing from 0.6906 to 0.6785 and carbon storage decreasing from 1.480 × 10⁹ t to 1.476 × 10⁹ t. (2) In the study area, significant synergistic effects existed between carbon storage and habitat quality, carbon storage and soil retention, carbon storage and water yield, habitat quality and soil retention, and soil retention and water yield. However, there was a significant trade-off relationship between habitat quality and water yield. (3) During the study period, ecosystem service trade-offs and synergy relationships in the Dongting Lake area were jointly influenced by natural factors and human activities. Ranked by the magnitude of driving factor influence, they were land use type, land use intensity, vegetation coverage, temperature, and nighttime light. [Conclusions] Synergies dominated the ecosystem services in the research region, and the influence of natural factors behind them was greater than that of human activities. These research conclusions offer a scientific foundation for the institutional construction of the ecological compensation mechanism in the Dongting Lake basin.

Counting the number of trees and obtaining information on tree crowns have always played important roles in the efficient and high-precision monitoring of forest resources. However, determining how to obtain the above information at a low cost and with high accuracy has always been a topic of great concern. Using deep learning methods to segment individual tree crowns in mixed broadleaf forests is a cost-effective approach to forest resource assessment. Existing crown segmentation algorithms primarily focus on discrete trees, with limited research on mixed broadleaf forests. The lack of datasets has resulted in poor segmentation performance, and occlusions in broadleaf forest images hinder accurate segmentation. To address these challenges, this study proposes a supervised segmentation method, SegcaNet, which can efficiently extract tree crowns from UAV images under natural light conditions. A dataset for dense mixed broadleaf forest crown segmentation is produced, containing 18,000 single-tree crown images and 1200 mixed broadleaf forest images. SegcaNet achieves superior segmentation results by incorporating a convolutional attention mechanism and a memory module. The experimental results indicate that SegcaNet’s $mIoU$ values surpass those of traditional algorithms. Compared with FCN, Deeplabv3, and MemoryNetV2, SegcaNet’s $mIoU$ is increased by 4.8%, 4.33%, and 2.13%, respectively. Additionally, it reduces instances of incorrect segmentation and over-segmentation.

In real-world wildfire monitoring, haze, glare, refraction artifacts, and related visual effects can cause false alarms and missed detections. This study explores two implementations: ContextFireVLM and ContextFireAgent. ContextFireVLM fine-tunes Llama-3.2 Vision with LoRA to improve the base model’s ability to identify and analyze wildfires; ContextFireAgent equips an agent with YOLO as a tool to improve the same capabilities. The first and second assessors independently inspect the full image, and the second assessor invokes YOLO only when it reports uncertainty; when their judgments agree, the final reviewer preserves the agreed decision. Because the two implementations use different models, inputs, and training pipelines, these results do not establish a causal comparison or general superiority between LoRA adaptation and the agent design.