首页 | 本学科首页   官方微博 | 高级检索  
相似文献
 共查询到20条相似文献,搜索用时 15 毫秒
1.
赤腹松鼠(Callosciurus erythralus)春季生境特征初步分析   总被引:1,自引:0,他引:1  
2009年2月至5月,在广西龙江河畔对赤腹松鼠(C.erythralus)的春季生境特征进行了分析.野外共测量了57个10m×10m样方中的13个生态因子,并运用频次分析和主成分分析的方法,对赤腹松鼠的春季生境选择因子进行了分析.结果表明,赤腹松鼠春季生境的主要特征为:郁闭度良好,水源距离<30m,坡度20~40°,避风性良好,坡向以东坡和南坡为主,坡位中坡位或上坡位,食物因子良好,人为干扰距离<10m,海拔50~100m,乔木密度<50株,乔木距离低于4m,灌木密度低于200株,灌木距离<2m.影响赤腹松鼠春季生境选择的主要因子为郁闭度、避风性、坡度、坡位和灌木距离;次要因子为海拔、人为干扰距离、乔木距离、水源距离、乔木密度、灌木密度、食物丰富度、坡向.  相似文献   

2.
马金超  郭振  许昊  宁焕杉  沈家伟  张志卫 《生态学报》2024,44(11):4770-4781
单木分割对于森林资源调查具有重要的意义,不同结构复杂度的森林单木分割算法的选择以及分割参数的选取对分割精度有着很大的影响。以山东田横岛为研究区,基于无人机正射影像与激光雷达数据,首先提取海岛森林典型植被二维与三维特征,然后利用随机森林算法对不同树种的树木进行分类,最后基于分类后的点云数据,选取不同结构复杂度的森林样地,对比分析聚类算法、层堆叠算法、分水岭算法在不同复杂度林区的适用性。结果表明:(1)随机森林算法结合单木二维、三维特征可有效对混交林树木进行分类,模型总体的精度为94.51%,Kappa系数为0.9038;(2)聚类算法对结构简单的林区具有更高的分割精度(F=96.41),但依赖于分割参数的选取;面对复杂单木集群,分水岭算法总体得分波动最小(ΔF=14.56),表现出较强的稳定性;(3)混交林预先进行树种分类可有效改善单木分割环境,相比于直接进行单木分割,聚类算法、层堆叠算法、分水岭算法的分割精度均得到不同程度的提升(ΔF1=10.06,ΔF2=9.51,ΔF3=12.6)。  相似文献   

3.
随机森林模型在分类与回归分析中的应用   总被引:25,自引:0,他引:25  
李欣海 《昆虫知识》2013,50(4):1190-1197
随机森林(random forest)模型是由Breiman和Cutler在2001年提出的一种基于分类树的算法。它通过对大量分类树的汇总提高了模型的预测精度,是取代神经网络等传统机器学习方法的新的模型。随机森林的运算速度很快,在处理大数据时表现优异。随机森林不需要顾虑一般回归分析面临的多元共线性的问题,不用做变量选择。现有的随机森林软件包给出了所有变量的重要性。另外,随机森林便于计算变量的非线性作用,而且可以体现变量间的交互作用(interaction)。它对离群值也不敏感。本文通过3个案例,分别介绍了随机森林在昆虫种类的判别分析、有无数据的分析(取代逻辑斯蒂回归)和回归分析上的应用。案例的数据格式和R语言代码可为研究随机森林在分类与回归分析中的应用提供参考。  相似文献   

4.
    
Nest-site availability limits cavity-using populations in many harvested forests; however, little is known about the extent of nest-site limitation in mature forests with a full complement of excavator species and intact processes of cavity creation and loss. To examine the role of nest-site availability in limiting cavity-using populations in mature mixed conifer forests in central British Columbia, Canada, we conducted an 11-year before-after control-impact experiment in which we increased nest-site availability via nest box addition. Our 7 sites (3 treatments, 4 controls) had low cavity densities (<2/ha) prior to treatment and cavity occupation rates were also low (<10%/yr), which is a relationship often cited in the literature as evidence of non-limitation in cavity-nesting populations. Following nest box addition at our treatment sites, which tripled the availability of cavities, total density of bird and mammal nests more than tripled. Density of mountain chickadee (Poecile gambeli) nests increased 9-fold on treatment sites and returned to pre-treatment levels following box removal, suggesting that chickadee populations were limited by cavity availability at our study sites. Density of red squirrel (Tamiasciurus hudsonicus) and northern flying squirrel (Glaucomys sabrinus) nests and roosts also increased significantly at treatment sites following box addition and declined following box removal. We noted little change in chickadee or squirrel nest density at control sites monitored concurrently. Squirrels preferred large-sized over small-sized boxes, and significantly enlarged the entrance areas of small boxes by chewing, suggesting that there may have been a shortage of suitable nest and roost sites for them in our study area. We contend that low cavity occupancy rates may not accurately reflect nest-site availability for cavity nesters in mature forests, and that cavity size may influence the true availability of cavities on the landscape. © 2011 The Wildlife Society.  相似文献   

5.
基于随机森林的不同径级树木死亡影响因子研究   总被引:1,自引:0,他引:1       下载免费PDF全文
树木死亡是多种因素共同作用的结果,然而不同径级的林木,导致其死亡的原因是否相同尚不明确。以坝上地区小叶杨为对象,于2016-2017年调查了100块20m×20m样地。以林木胸径、林木相对断面积、林龄、林分密度、不同土层土壤含水量和土壤容重等11个因子为输入变量,运用随机森林算法分别不同径级建立树木死亡模型,分析各因子对不同径级树木死亡的影响大小,并得到关键影响因子的阈值。结果表明:依据不同径级树木死亡与影响因子的相关性不同,可将林木径级划分为 < 10cm、10-25cm、 > 25cm;胸径 < 10cm时,林木相对断面积对树木死亡的影响最大,当0.23 < 林木相对断面积 < 0.43时,树木死亡率最低;胸径为10-25cm时,密度对树木死亡的影响最大,当密度 > 600株/hm2时,树木死亡率随着密度的增加明显上升;胸径 > 25cm时,20-60cm土壤水分和林分断面积对树木死亡的影响最大且相近,当20-60cm土壤质量含水量 > 5%时,树木死亡率明显下降;基于随机森林算法建立的树木死亡模型可以解释树木死亡70%-90%的变异,具有一定的预测优势和应用价值,可为坝上地区树木死亡预测和防护林的经营管护提供参考。  相似文献   

6.
7.
基于不同决策树的面向对象林区遥感影像分类比较   总被引:1,自引:0,他引:1  
陈丽萍  孙玉军 《生态学杂志》2018,29(12):3995-4003
面向地理对象影像分析技术(GEOBIA)是影像分辨率越来越高的背景下的产物.如何提高高分辨率影像分类精度和分类效率是影像处理的重要议题之一.本研究对QuickBird影像多尺度分割后的对象进行分类,分析了C5.0、C4.5、CART决策树算法在林区面向对象分类中的效率,并与kNN算法的分类精度进行比较.利用eCognition软件对遥感影像进行多尺度分割,分析得到最佳尺度为90和40.在90尺度下分离出植被和非植被后,在40尺度下提取不同类别植被的光谱、纹理、形状等共21个特征,并利用C5.0、C4.5、CART决策树算法分别对其进行知识挖掘,自动建立分类规则.最后利用建立的分类规则分别对植被区域进行分类,并比较分析其精度.结果表明: 基于决策树的分类精度均高于传统的kNN法.其中,C5.0方法的精度最高,其总体分类精度为90.0%,Kappa系数0.87.决策树算法能有效提高林区树种分类精度,且C5.0决策树的Boosting算法对该分类效果具有最明显的提升.  相似文献   

8.
    
Correlations between surface behavior and concurrent underwater vocalizations were modeled for common dolphins (Delphinus spp.) in the Southern California Bight (SCB) over multiple field seasons. Clicks, pulsed calls, and whistles were examined, with a total of 50 call features identified. Call features were used to classify behavior using random forest decision trees, with rates of correct classification reaching 80.6% for fast travel, 84.6% for moderate travel, 59.8% for slow travel, and 58% for foraging behavior. Common dolphins spent most of their time traveling. The highest number of clicks, pulsed calls, and complex whistles were produced during fast travel. In contrast, during foraging there were few pulsed calls and whistles produced, and the whistles were simple with narrow bandwidths and few harmonics. Behavior and vocalization patterns suggest nocturnal foraging in offshore waters as the primary feeding strategy. Group size and spacing were strongly correlated with behavior and rates of calling, with higher call rates in dispersed traveling groups and lower call rates in loosely aggregated foraging groups. These results demonstrate that surface behavior can be classified using vocalization data, which builds the framework for behavioral studies of common dolphins using passive acoustic monitoring techniques.  相似文献   

9.
10.
    
  1. Adenovirus is an emerging threat to red squirrel Sciurus vulgaris conservation, but confirming clinically significant adenovirus infections in red squirrels is challenging. Rapid intestinal autolysis after death in wild animals frequently obscures pathology characteristic of the disease in animals found dead.
  2. We review the available literature to determine current understanding of both subclinical and clinically significant adenovirus infections in free‐living wild and captive red squirrel populations.
  3. Benefits of scientific testing for adenovirus incorporating both transmission electron microscopy (TEM) and polymerase chain reaction (PCR) technologies are compared and contrasted. We favour viral particle detection using TEM in animals exhibiting enteropathy at post‐mortem and the use of PCR to detect subclinical cases where no enteric abnormalities are observed.
  4. Adenoviral infections associated with re‐introduction studies are evaluated by examination of sporadic cases in wild populations and of data from captive collections used to service such studies.
  5. The paucity of data available on adenovirus infection in grey squirrel Sciurus carolinensis populations is documented, and we highlight that although subclinical virus presence is recorded in several locations in Great Britain and in Italy, no clinically significant disease cases have been detected in the species thus far.
  6. Current speculation about potential interspecific infection between sciurids and other woodland rodents such as wood mice Apodemus sylvaticus is examined. Where subclinical adenovirus presence has been detected in sympatric populations using the same point food sources, husbandry methods may be used to diminish the potential for cross‐infection.
  7. Our findings highlight the importance of controlling disease in red squirrel populations by using clearly defined scientific methods. In addition, we propose hypothetical conservation benefits of restricting contact rates between red squirrels and sympatric grey squirrels and of limiting competition from other woodland rodent species.
  相似文献   

11.
通常来讲,生态学者对于解释生态关系、描述格局和过程、进行空间或时间预测比较感兴趣。这些工作可以通过模拟输出值(响应)与一些特征值(即解释变量)的关系来实现。然而,生态数据模拟遇到了挑战,这是因为响应变量和预测变量可能是连续变量或离散变量。需要解释的生态关系通常是非线性的,并且解释变量之间具有复杂的相互作用关系。响应变量和解释变量存在缺失值并不是不常有的现象,奇异值也经常出现在生态数据中。此外,生态学者通常希望生态模型即要易于建立又易要于解释。通常是利用多种统计方法来分析处理各种各样情景中出现的独特的生态问题,这些模型包括(多元)逻辑回归、线性模型、生存模型、方差分析等等。随机森林是一个可以处理所有这些问题的有效方法。随机森林可以用来做分类、聚类、回归和生存分析、评估变量的重要性、检测数据中的奇异值、对缺失数据进行插补等。鉴于随机森林本身在算法上的优势,将就随机森林在生态学中的应用进行总结,对建模过程进行概述,并以云南松分布模拟研究为例,对其主要功能特点进行案例展示。通过对随机森林的一般术语、概念和建模思想进行介绍,有利于读者掌握本方法的应用本质,可以预见随机森林在生态学研究中将得到更多的应用和发展。  相似文献   

12.
13.
UsingEcopath, a trophic mass-balance modeling framework, we developed six models of a Canadian boreal forest food web centered around snowshoe hares, which have conspicuous 10-year population cycles. Detailed models of four phases of the cycle were parameterized with long-term population data for 12 vertebrate taxa. We also developed five other models that, instead of observed data, used parameter values derived from standard assumptions. Specifically, in the basic model, production was assumed to equal adult mortality, feeding rates were assumed to be allometric, and biomass was assumed to be constant. In the actual production, functional response, and biomass change models, each of these assumed values from the basic model was replaced individually by field data. Finally, constant biomass models included actual production by all species and functional responses of mammalian predators and revealed the proportion of herbivore production used by species at higher trophic levels. By comparing these models, we show that detailed information on densities and demographics was crucial to constructing models that captured dynamic aspects of the food web. These detailed models reinforced an emerging picture of the causes and consequences of the snowshoe hare cycle. The snowshoe hare decline and low phases were coincident with times when per capita production was relatively low and predation pressure high. At these times, ecotrophic efficiencies (EE) suggest there was little production that remained unconsumed by predators. The importance of both production and consumption implies that bottom–up and top–down factors interacted to cause the cycle. EEs of other herbivores (ground squirrels, red squirrels, small mammals, small birds, grouse) were generally low, suggesting weak top–down effects. Predation rates on these “alternative” prey, except ground squirrels, were highest when predators were abundant, not when hares were rare; consequently, any top–down effects reflected predator biomass and were not a function of diet composition or functional responses. Finally, several predators (lynx, coyotes, great-horned owls) showed clear bottom–up regulation, reproducing only when prey exceeded threshold densities. Taken altogether, these results demonstrate that ecosystem models parameterized by population data can describe the dynamics of nonequilibrial systems, but only when detailed information is available for the species modeled. Received 30 November 2000; Accepted 6 September 2001.  相似文献   

14.
Pueppke  S. G. 《Plant and Soil》1988,109(2):189-193
Current and one-year-old foliage was collected from sixty-five red spruce trees growing in thirteen stands at different elevations in the Green Mountains of Vermont and Adirondacks of New York. Sample trees were randomly selected from visually healthy trees at each site. Foliage was analyzed for major and minor elements. In July 1984, foliar Ca, Mg, and Zn concentrations were significantly greater at low than at high elevations. In October 1984, Ca, Mg, and Zn concentrations were higher at low elevations and Ca and Mg concentrations varied significantly among locations within elevational groups. Nitrogen concentration was significantly higher in the high-elevation group in July but not in October. The average red spruce foliar Mg concentration at the end of the growing season in the high elevation stands (442 mg kg−1) is much lower than values reported for other mature red spruce stands in the eastern United States.  相似文献   

15.
    
Although open-cup nesting birds generally face increased risk of nest depredation from forest edge predators and brood parasites in fragmented temperate landscapes, little information exists to assess such risks in tropical birds. We compared nesting success of real birds' nests in large and small forest fragments to a control site in Caribbean lowland wet forest of Costa Rica. Pooling across species, nesting success was significantly greater in unfragmented forest than in either small, isolated fragments or the La Selva Biological Reserve, which is at the tip of a forest 'peninsula' embedded in a largely deforested landscape. Nesting success in isolated fragments did not vary according to distance from edge, suggesting that predators in fragments act throughout these forest patches. The case for increased nest predation as a plausible mechanism to explain the documented decline of forest interior bird populations in this fragmented tropical landscape is enhanced by a simple demographic model that suggests nesting success is likely too low to maintain populations at La Selva and in the fragments. The fact that the large (> 1000 ha) La Selva forest reserve is experiencing nest predation rates similar to those in much smaller fragments is cause for concern. Our results make a strong case for additional studies to document the identities of nest predators in both fragmented and unfragmented forests in such tropical forest landscapes.  相似文献   

16.
    
Captive and field studies both provide valuable and complementary information that lead to a better understanding of a species' behavioral ecology. Here, we review studies from wild, captive, and semi-free ranging populations of squirrel monkeys (Saimiri sp.), in order to (a) provide a more current (1985-2010) review of Saimiri behavioral ecology and (b) illustrate that integrating data collected in a variety of settings is an effective approach to addressing ecological questions in primates. Captive environments, such as zoological facilities and research colonies, can be advantageous to researchers by allowing longitudinal studies of behavior and reproduction, as well as providing opportunities for gathering data on life history, because physiological and life history data are known for individual animals. Studies of field populations can provide contextual information regarding the adaptive nature of behaviors that are studied in captivity. Squirrel monkeys are small, neotropical primates that have extensively been used in captive research. As the last in-depth review of Saimiri biology was published in 1985 [Rosenblum & Coe, The squirrel monkey. New York: Academic Press], we review studies since conducted on Saimiri ecology, life history, social behavior, reproduction, and conservation. Our review indicates that there is much variation in socioecology and life history traits between Saimiri species and, surprisingly, also between populations of the same species studied at different locales. In addition, much is known about squirrel monkey reproductive physiology, basic ecology, and vocal communication, but data are still lacking in the fields of life history and some adaptive components and social behavior. In particular, longitudinal studies in the field would be particularly relevant for a genus with a slow life history such as Saimiri. Finally, few data (captive or wild) are available on S. ustus and S. vanzolinii, though at least one of these species is threatened.  相似文献   

17.
    
  1. Bioacoustic methods play an increasingly important role for the detection of insects in a range of surveillance and monitoring programmes.
  2. Weak-flying insects evade detection because they do not yield sufficient audio information to capture wingbeat and harmonic frequencies. These inaudible insects often pose a significant threat to food security as pests of key agricultural crops worldwide.
  3. Automatic detection of such insects is crucial to the future of crop protection by providing critical information to assess the risk to a crop and the need for preventative measures.
  4. We describe an experimental set-up designed to derive audio recordings from a range of weak-flying aphids and beetles using an LED array.
  5. A rigorous data processing pipeline was developed to extract meaningful features, linked to morphological characteristics, from the audio and harmonic series for six aphid and two beetle species.
  6. An ensemble of over 50 bioacoustic parameters was used to achieve species discrimination with a success rate of 80%. The inclusion of the dominant and fundamental frequencies improved prediction between beetles and aphids because of large differences in wingbeat frequencies.
  7. At the species level, error rates were minimized when harmonic features were supplemented by features indicative of differences in species' flight energies.
  相似文献   

18.
  总被引:5,自引:2,他引:3  
Although a parapoxvirus harmful to red squirrels is present in UK squirrel populations it has not been considered a major cause of red squirrel decline, and replacement by the introduced grey squirrel, mainly because diseased individuals are rarely observed. By developing a generic model we show that parapoxvirus is likely to have played a crucial role in the red squirrel decline even though the prevalence of infection is low. Conservationists are quite rightly concerned with the invasion of exotic organisms such as the grey squirrel. Our work emphasizes that they, along with other ecologists, should pay particular attention to pathogens, even when they occur at low prevalence.  相似文献   

19.
    
Plasma concentrations of progesterone (P4) and 17β-oestradiol (E2) in juvenile, pre-ovulatory, early, mid- or late pregnancy stages of female blue sharks Prionace glauca were analysed. Concentrations of P4 were significantly higher in pregnant than in non-pregnant individuals, whereas E2 concentrations increased with embryonic and follicular development. A highly accurate (86.1%) random forest classification model was developed to predict shark pregnancy. It is proposed that hormone concentrations could be used for the subsequent non-lethal determination of female P. glauca reproductive state.  相似文献   

20.
设为首页 | 免责声明 | 关于勤云 | 加入收藏

Copyright©北京勤云科技发展有限公司  京ICP备09084417号