Objective: The paper aims to analyze the dynamic characteristics of litter production and nutrient return of the forest ecosystems in subtropical areas, and provide a theoretical basis for the nutrient cycling study i...Objective: The paper aims to analyze the dynamic characteristics of litter production and nutrient return of the forest ecosystems in subtropical areas, and provide a theoretical basis for the nutrient cycling study in southwest Hubei Province and carbon sink function of the whole forest ecosystem. Methods: Three typical forest stands (Chinese fir plantation, Cryptomeria fortunei plantation and evergreen and deciduous broad-leaved mixed forest) in Golden Mountain Forest Farm in southwest Hubei Province were investigated and monitored continuously for the litter types and productivity and nutrient return. Results: The annual litter productivity of the three forest stands ranged from 161.77 to 396.26 kg·hm<sup>-2</sup>;Litters of branches, leaves and reproductive organs accounted for 14.14% - 20.85%, 33.26% - 78.33%, 7.52% - 42.18% of the total, respectively;The litter productivity and total litter productivity of each composition in the three forest stands show unimodal or bimodal changes over months, and the total litter productivity reached the highest value in January, April and October respectively. For different nutrient contents of the three forest stands, the common feature is C > N. The order of nutrient return amount from greatest to least is evergreen and deciduous broad-leaved mixed forest, Cryptomeria fortunei plantation and Chinese fir plantation. For different nutrient return amounts, the common feature is C > N, and the nutrient return amounts are 76.51-180.69 kg·hm<sup>-2</sup> and 2.3 - 5.71 kg·hm<sup>-2</sup> respectively. Conclusion: The annual litter productivity and nutrient return amount of the evergreen and deciduous broad-leaved mixed forest are the highest among the three forest stands. Therefore, protecting the evergreen and deciduous broad-leaved mixed forest and studying the litter changes of Chinese fir plantation and Cryptomeria fortunei plantation are of far-reaching significance for the development of sustainable forest management in this region and the further improvement of the carbon sequestration function of the whole forest ecosystem.展开更多
文摘Objective: The paper aims to analyze the dynamic characteristics of litter production and nutrient return of the forest ecosystems in subtropical areas, and provide a theoretical basis for the nutrient cycling study in southwest Hubei Province and carbon sink function of the whole forest ecosystem. Methods: Three typical forest stands (Chinese fir plantation, Cryptomeria fortunei plantation and evergreen and deciduous broad-leaved mixed forest) in Golden Mountain Forest Farm in southwest Hubei Province were investigated and monitored continuously for the litter types and productivity and nutrient return. Results: The annual litter productivity of the three forest stands ranged from 161.77 to 396.26 kg·hm<sup>-2</sup>;Litters of branches, leaves and reproductive organs accounted for 14.14% - 20.85%, 33.26% - 78.33%, 7.52% - 42.18% of the total, respectively;The litter productivity and total litter productivity of each composition in the three forest stands show unimodal or bimodal changes over months, and the total litter productivity reached the highest value in January, April and October respectively. For different nutrient contents of the three forest stands, the common feature is C > N. The order of nutrient return amount from greatest to least is evergreen and deciduous broad-leaved mixed forest, Cryptomeria fortunei plantation and Chinese fir plantation. For different nutrient return amounts, the common feature is C > N, and the nutrient return amounts are 76.51-180.69 kg·hm<sup>-2</sup> and 2.3 - 5.71 kg·hm<sup>-2</sup> respectively. Conclusion: The annual litter productivity and nutrient return amount of the evergreen and deciduous broad-leaved mixed forest are the highest among the three forest stands. Therefore, protecting the evergreen and deciduous broad-leaved mixed forest and studying the litter changes of Chinese fir plantation and Cryptomeria fortunei plantation are of far-reaching significance for the development of sustainable forest management in this region and the further improvement of the carbon sequestration function of the whole forest ecosystem.
文摘为构建高精度的森林地表凋落物含水率预测模型,以黑龙江省哈尔滨市3种典型林分白桦林(Betula platyphylla)、蒙古栎林(Quercus mongolica)和兴安落叶松林(Larix gmelini)为研究对象,采用自制的便携式气象因子检测仪进行实时检测,通过随机森林(Random Forest,RF)、反向传播(Back Propagation,BP)和长短期记忆网络(Long Short Term Memory,LSTM)3种代表性的机器学习方式对3种林分分别构建凋落物含水率预测模型。研究表明,不同的林分在不同的机器学习方式下构建的预测模型的预测精度均不同;在不同的机器学习构建的模型中3种林分的预测精度均为LSTM>BP>RF;LSTM在地表凋落物含水率预测模型的研究中效果最优。此研究结果说明LSTM在地表凋落物含水率的预测中表现突出,对提高基于气象要素法预测地表凋落物含水率模型精度具有重要的理论意义和实践指导价值。