基于RF-LSTM的鸡舍恶臭气体预测研究

以鸡舍氨气为研究对象,对鸡舍氨气预测模型进行了研究.首先,利用随机森林算法（RF）对影响鸡舍氨气浓度的环境变量进行重要性排序,选取温度、湿度、光照、气象温度、降雨量作为模型的输入变量;在此基础上,构建了基于长短时记忆神经网络（LSTM）的鸡舍氨气浓度预测模型,并将提出的预测模型应用于江苏省宜兴市某养鸡场的氨气浓度预测中,并与LSTM模型、RF-Elman模型和RF-BP模型进行了对比实验,结果表明,基于RF-LSTM模型的预测效果最好,其平均绝对误差（MAE）、平均绝对百分误差（MAPE）和均方根误差（RMSE）分别为0.9183、4.9637%和1.4262;同时,为了验证该模型的性能,本文还实现了不同时间尺度的鸡舍氨气浓度预测,提前2h、3h、4h、5h氨气预测的平均绝对误差（MAE）分别为1.6218、2.1991、2.8553和3.0677.本文提出的预测模型提高了鸡舍氨气浓度的预测精度,可为减少鸡舍恶臭气体排放提供科学依据.     

未来气候变化对武夷山自然保护区毛竹异戊二烯排放速率的影响

采用全球气候模式Nor ESM1-M产生的RCP2.6、RCP4.5、RCP6.0和RCP8.5气候变化情景数据和植物异戊二烯排放计算模型,模拟分析了未来气候变化对武夷山自然保护区毛竹(Phyllostachys pubescens)异戊二烯排放速率的影响.结果显示,气候变化下武夷山自然保护区气温上升,年降水量和辐射强度波动较大,呈增加或下降趋势.毛竹异戊二烯平均日排放速率在未来气候变化情景下比基准情景下高约30μg·g~(-1)·d~(-1),在RCP8.5情景下比基准情景下高约48μg·g~(-1)·d~(-1);毛竹异戊二烯日排放速率在未来气候变化情景与基准情景下的差异在1~90 d和301~365 d较小,在91~300 d差异较大;相比基准情景,未来气候变化情景下毛竹异戊二烯日排放速率在1~190 d(平均增加15%以上)和271~365 d(平均增加20%)增幅较大,在191~270 d增幅较小,在RCP8.5情景下增幅最大(平均增加17%).另外,毛竹异戊二烯年排放速率在未来气候变化情景下比基准情景下约高10000μg·g~(-1)·a~(-1)以上,在RCP8.5情景下比基准情景下约高13%.研究表明,未来气候变化将使毛竹异戊二烯排放速率增加.     

基于高分1号影像的森林植被信息提取

实时最新森林植被信息的提取是林业航空植保作业的必要前提。论文以安徽省蚌埠市为研究区域,探讨了基于高分1号卫星遥感数据在亚热带农林植被混合地区的森林植被信息提取。根据植被物候信息差异选择了提取森林植被信息的5个关键时期高分影像,采用分区决策树方法监测森林植被的空间分布和面积信息,并与未分区决策树法的提取结果进行比较。结果表明：采用分区决策树法和未分区决策树法对于大中尺度森林植被信息提取的总体精度均优于85%。但分区决策树森林植被提取总体精度达到90.72%,较未分区决策树法提高3.80%、4.65%,Kappa系数达到0.81,较未分区决策树法提高约0.07~0.10,结合植被物候信息的分区决策树森林植被提取法好于未分区决策树法,能够满足林业航空植保作业的精度需求。具有较高空间分辨率、宽覆盖、短重访周期的高分1号影像,对于大区域的林业航空植保当年最新森林植被信息的提取表现出较大的潜力。     

气候变化和人类活动对武江流域年径流及最大日流量影响的定量分析

针对气候变化与人类活动对流域年径流及最大日流量变化影响的定量识别问题,以华南湿润区武江流域为例,分别采用HIMS（Hydro-Informatic Modeling System）模型和敏感性系数法,从日和年尺度定量模拟和评估气候变化与人类活动对流域年最大日流量和径流变化的影响过程及贡献率。结果表明：HIMS模型在武江流域适用性良好,日尺度模型率定期和验证期的纳西效率系数分别为0.85和0.77,水量平衡误差绝对值分别为3.1%和3.3%;两种方法均表明气候变化是引起流域年径流量增加的主要因素,人类活动导致了流域径流量的减少,但贡献率较小。气候变化与人类活动导致了流域年最大日流量的增加,气候变化对年最大日流量增加的贡献率为94%,而人类活动的贡献率则为6%。相较于年均径流量,气候变化对年最大日流量的影响更为显著。     

长江源区水文气象要素变化及其与大尺度环流因子关系研究

长江源区水文气象要素变化及其归因研究一直备受全球关注,现阶段研究多侧重于水文气象要素时空变化特征分析,针对长江源区水文气象要素与大尺度环流因子相互关系的研究不足。论文利用Mann-Kendall法、去趋势波动分析法和小波分析法,探究长江源区1957—2012年水文气象要素趋势性、波动性和周期性变化规律,分析水文气象要素与大尺度环流因子的相关关系,通过研究水汽通量揭示大尺度环流因子对水文气象要素变化的驱动机制。结果表明：20世纪90年代,长江源区气候暖干化,进入21世纪后,长江源区气候暖湿化趋势明显;长江源区水文气象要素序列具有正长程相关性,长江源区气候未来会继续呈现暖湿化变化趋势。长江源区水文气象要素都存在着1~5、10~24和25~45 a三种时间尺度周期变化规律。南亚季风是影响长江源区降水量和流量较为重要的大气环流因子,南亚季风驱动下的西南方向气流是长江源区主导气流和水汽来源。     

Floristics, performance and prognosis of historical trees in the urban forest of Guangzhou City (China)

Outstanding historical trees embedded in cities constitute pertinent environmental assets, yet they are widely threatened in third-world cities. Inadequate understanding of this valuable natural-cum-cultural heritage hinders proper conservation. A case study of Guangzhou in south China evaluated floristic composition, age profile and biomass structure of historical trees, assessed their performance in major habitats (institutional, park and roadside), and established a prognosis for future growth and management. The 348 historical trees examined belonged to only 25 species, vis-à -vis 254 trees in the entire urban forest, dominated by five species and native members. Roadside had more trees, followed by institutional and park, with merely the most common four species shared by all habitats. The limited commonality reflected tree-performance differentiation by habitats exerting selection pressure on species. The institutional growth-regime was more conducive to nurturing high-caliber specimens, whereas park is less capable. Individual species achievement by habitats, derived from tree-count ranking and relative-abundance indices, could inform species choice and tree conservation. Few trees exceeded 300 years of age in the millennium-old city, echoing a history of intense tree—city conflicts. Potential life-span, trunk and crown diameters indicated ample opportunities for further expansion of biomass and landscape impacts, which would be straitjacketed by the tightening urban fabric.     

阿图什市空气质量变化趋势分析

分析了阿图什市“八五”至“九五”期间空气质量监测数据不同年度、不同季节及不同污染因子的动态变化趋势．结合当地能源结构、气候特征、城市环境综合发展水平，指出影响阿图什市空气质量的主要因素，为防治和减轻阿图什市的空气污染提供了科学依据。     

Sharing the geo-Referenced Results of Climate Change Impact Research

The Prairie Adaptation Research Collaborative (PARC) has implemented an Internet Map Server (IMS) at the PARC web site (www.parc.ca) to 1) disseminate the geo-referenced results of PARC sponsored research on climate change impacts and adaptation, and 2) address data, information and knowledge management within the PARC network of researchers and partners. PARC facilitates interdisciplinary research on adaptation to the impacts of climate change in the Canadian Prairie Provinces. The web site is intended as a platform for sharing information and encouraging discussion of climate change impacts and adaptation. The IMS enables scientists and stakeholders to apply simple climate change scenarios to geo-referenced biophysical and social data, and dynamically create maps that display the geographic distribution of potential impacts of climate change. With a limited capacity for spatial analysis, most geo-processing and the climate impact modeling is done offline within a GIS environment. The IMS will serve the output from climate impact models, such that the model results can be customized by the web site user and be most readily applied to the planning and analysis of adaptation strategies.     

Assessing protected area effectiveness using surrounding (buffer) areas environmentally similar to the target area

Many studies are based on the assumption that an area and its surrounding (buffer) area present similar environmental conditions and can be compared. For example, in order to assess the effectiveness of a protected area, the land use/cover changes are compared inside the park with its surroundings. However, the heterogeneity in spatial variables can bias this assessment: we have shown that most of the protected areas in Mexico present significant environmental differences between their interior and their surroundings. Therefore, a comparison that aims at assessing the effectiveness of conservation strategies, must be cautioned. In this paper, a simple method which allows the generation of a buffer area that presents similar conditions with respect to a set of environmental variables is presented. The method was used in order to assess the effectiveness of the Calakmul Biosphere Reserve, a protected area located in the south-eastern part of Mexico. The annual rate of deforestation inside the protected area, the standard buffer area (based upon distance from the protected area only) and the similar buffer area (taking into account distance along with some environmental variables) were 0.3, 1.3 and 0.6%, respectively. These results showed that the protected area was effective in preventing land clearing, but that the comparison with the standard buffer area gave an over-optimistic vision of its effectiveness.     

Status and Trends of Dissolved Oxygen in Corpus Christi Bay, Texas, U.S.A.

The purpose of this studywas to determine status and long-term trends of dissolved oxygen concentrations (DO) in Corpus Christi Bay, Texas, U.S.A. A 20-year record of randomized stations was used to determine the trend of surface water DO, salinity, and temperature over space and time. A 13-year record of two fixed stations was used to determine the temporal nutrient trends. A 10-year record of fixed stations in the southeastern region of Corpus Christi Bay was used to determine the status of disturbance caused by low DO in bottom waters. From 1982 to 2002, there was a significant decrease in surface water DO at a rate of 0.06 mg L⁻¹ yr⁻¹ and a significant increase in surface water temperature at a rate of 0.07°C yr⁻¹. The southeastern region of Corpus Christi Bay had the lowest average DO, and during July and August, DO are steadily declining at a rate of 0.09 mg L⁻¹ yr⁻¹. It is not likely that eutrophication is causing hypoxia, because freshwater inflow rates have significantly decreased since 1941 and nutrient levels have not changed from 1987 to 2000. Even though long-term trends indicate that average surface DO is decreasing, disturbance by hypoxia appears to be stable, but this may be due to just eight years of data. In fact, if the current trend continues, surface water DO will not meet exceptional aquatic life standards (≤5 mgL⁻¹) in 2032.