Characterizing lentic freshwater fish assemblages using multiple sampling methods

Characterizing fish assemblages in lentic ecosystems is difficult, and multiple sampling methods are almost always necessary to gain reliable estimates of indices such as species richness. However, most research focused on lentic fish sampling methodology has targeted recreationally important species, and little to no information is available regarding the influence of multiple methods and timing (i.e., temporal variation) on characterizing entire fish assemblages. Therefore, six lakes and impoundments (48–1,557 ha surface area) were sampled seasonally with seven gear types to evaluate the combined influence of sampling methods and timing on the number of species and individuals sampled. Probabilities of detection for species indicated strong selectivities and seasonal trends that provide guidance on optimal seasons to use gears when targeting multiple species. The evaluation of species richness and number of individuals sampled using multiple gear combinations demonstrated that appreciable benefits over relatively few gears (e.g., to four) used in optimal seasons were not present. Specifically, over 90 % of the species encountered with all gear types and season combinations (N?=?19) from six lakes and reservoirs were sampled with nighttime boat electrofishing in the fall and benthic trawling, modified-fyke, and mini-fyke netting during the summer. Our results indicated that the characterization of lentic fish assemblages was highly influenced by the selection of sampling gears and seasons, but did not appear to be influenced by waterbody type (i.e., natural lake, impoundment). The standardization of data collected with multiple methods and seasons to account for bias is imperative to monitoring of lentic ecosystems and will provide researchers with increased reliability in their interpretations and decisions made using information on lentic fish assemblages.     

Pattern Classification and Recognition of Invertebrate Functional Groups Using Self-Organizing Neural Networks

Self-organizing neural networks can be used to mimic non-linear systems. The main objective of this study is to make pattern classification and recognition on sampling information using two self-organizing neural network models. Invertebrate functional groups sampled in the irrigated rice field were classified and recognized using one-dimensional self-organizing map and self-organizing competitive learning neural networks. Comparisons between neural network models, distance (similarity) measures, and number of neurons were conducted. The results showed that self-organizing map and self-organizing competitive learning neural network models were effective in pattern classification and recognition of sampling information. Overall the performance of one-dimensional self-organizing map neural network was better than self-organizing competitive learning neural network. The number of neurons could determine the number of classes in the classification. Different neural network models with various distance (similarity) measures yielded similar classifications. Some differences, dependent upon the specific network structure, would be found. The pattern of an unrecognized functional group was recognized with the self-organizing neural network. A relative consistent classification indicated that the following invertebrate functional groups, terrestrial blood sucker; terrestrial flyer; tourist (nonpredatory species with no known functional role other than as prey in ecosystem); gall former; collector (gather, deposit feeder); predator and parasitoid; leaf miner; idiobiont (acarine ectoparasitoid), were classified into the same group, and the following invertebrate functional groups, external plant feeder; terrestrial crawler, walker, jumper or hunter; neustonic (water surface) swimmer (semi-aquatic), were classified into another group. It was concluded that reliable conclusions could be drawn from comparisons of different neural network models that use different distance (similarity) measures. Results with the larger consistency will be more reliable.     

Comparison between electrofishing and snorkeling surveys to describe fish assemblages in Laurentian streams

We evaluated the relative performance of electrofishing and visual surveys (snorkeling) for estimating the abundance of combinations of fish species and size classes in rivers. We also assessed the effect of environmental conditions on potential differences between the results obtained using these two sampling methods. Sampling sites were distributed in the Laurentian region of Québec. Both methods were used while sections were blocked. Three snorkelers swam the river sections upstream while identifying and counting fish of each species and size classes. Three-pass electrofishing was performed in the same sites and abundances were estimated with a maximum likelihood depletion model. Greater abundances of fish were observed by snorkeling than by electrofishing at all sites. Snorkeling species richness was higher or equal to electrofishing richness in, respectively, 60 % and 40 % of sampled sites. Differences in the fish communities observed by both sampling methods were not related to environmental conditions. The results of our work are therefore contrary to that of most published studies that suggested the use of electrofishing over visual surveys. This study highlights that conclusions derived from previous work on sampling gear comparisons may not be generalisable; rather survey methods might benefit from being selected based on fish community composition.     

Using Multi-Scale Sampling and Spatial Cross-Correlation to Investigate Patterns of Plant Species Richness

Land managers need better techniques to assess exoticplant invasions. We used the cross-correlationstatistic, I _YZ, to test for the presence ofspatial cross-correlation between pair-wisecombinations of soil characteristics, topographicvariables, plant species richness, and cover ofvascular plants in a 754 ha study site in RockyMountain National Park, Colorado, U.S.A. Using 25 largeplots (1000 m²) in five vegetation types, 8 of 12variables showed significant spatial cross-correlationwith at least one other variable, while 6 of 12variables showed significant spatial auto-correlation. Elevation and slope showed significant spatialcross-correlation with all variables except percentcover of native and exotic species. Percent cover ofnative species had significant spatialcross-correlations with soil variables, but not withexotic species. This was probably because of thepatchy distributions of vegetation types in the studyarea. At a finer resolution, using data from ten1 m² subplots within each of the 1000 m² plots, allvariables showed significant spatial auto- andcross-correlation. Large-plot sampling was moreaffected by topographic factors than speciesdistribution patterns, while with finer resolutionsampling, the opposite was true. However, thestatistically and biologically significant spatialcorrelation of native and exotic species could only bedetected with finer resolution sampling. We foundexotic plant species invading areas with high nativeplant richness and cover, and in fertile soils high innitrogen, silt, and clay. Spatial auto- andcross-correlation statistics, along with theintegration of remotely sensed data and geographicinformation systems, are powerful new tools forevaluating the patterns and distribution of native andexotic plant species in relation to landscape structure.     

应用鱼类完整性指数F-IBI评价巢湖流域的主要河流健康

保护河流健康是河流生态管理的重要目标,鱼类对维护河流生态系统健康的作用重要且不可替代。依据2013年10月巢湖53个河流采样点的鱼类调查数据,采用假设参照值法构建鱼类生物完整性指数(F-IBI)的参照系统,进而评价河流健康状况。基于29个候选生物参数,通过参数与环境因子间的相关性分析、参数间的冗余分析及其分布范围分析,确定F-IBI由鱼类总物种数、鲤科鱼类物种数百分比、中下层鱼类物种数百分比、肉食性鱼类数量百分比、产粘性卵鱼类数量百分比和耐受性鱼类数量百分比等6个参数构成。利用比值法统一各参数量纲,F-IBI值即为各构成参数比值的累加,进而评价河流水生态健康的等级,其中亚健康的采样点占16.98%,健康状况一般的采样点占50.94%,健康状况较差的采样点占32.08%。此外,影响巢湖主要河流F-IBI的重要水环境变量是叶绿素a。对重点污染区域,修复水生态系统和制定科学合理的策略,是恢复巢湖流域生态健康的重要手段。     

The Influence of Linear Elements on Plant Species Diversity of Mediterranean Rural Landscapes: Assessment of Different Indices and Statistical Approaches

This paper mainly aims to study the linear element influence on the estimation of vascular plant species diversity in five Mediterranean landscapes modeled as land cover patch mosaics. These landscapes have several core habitats and a different set of linear elements -habitat edges or ecotones, roads or railways, rivers, streams and hedgerows on farm land- whose plant composition were examined. Secondly, it aims to check plant diversity estimation in Mediterranean landscapes using parametric and non-parametric procedures, with two indices: Species richness and Shannon index.Land cover types and landscape linear elements were identified from aerial photographs. Their spatial information was processed using GIS techniques. Field plots were selected using a stratified sampling design according to relieve and tree density of each habitat type. A 50×20 m² multi-scale sampling plot was designed for the core habitats and across the main landscape linear elements. Richness and diversity of plant species were estimated by comparing the observed field data to ICE (Incidence-based Coverage Estimator) and ACE (Abundance-based Coverage Estimator) non-parametric estimators.The species density, percentage of unique species, and alpha diversity per plot were significantly higher (p < 0.05) in linear elements than in core habitats. ICE estimate of number of species was 32% higher than of ACE estimate, which did not differ significantly from the observed values. Accumulated species richness in core habitats together with linear elements, were significantly higher than those recorded only in the core habitats in all the landscapes. Conversely, Shannon diversity index did not show significant differences.     

Fish abundances in shoreline habitats and submerged aquatic vegetation in a tidal freshwater embayment of the Potomac River

Submerged aquatic vegetation (SAV) is considered an important habitat for juvenile and small forage fish species, but many long-term recruitment surveys do not effectively monitor fish communities in SAV. To better understand the impact of recent large increases of SAV on the fish community in tidal freshwater reaches of the Potomac River, we compared traditional seine sampling from shore with drop ring sampling of SAV beds (primarily Hydrilla) in a shallow water (depths, <1.5 m) embayment, Gunston Cove. To accomplish this, we developed species-specific catch efficiency values for the seine gear and calculated area-based density in both shoreline and SAV habitats in late summer of three different years (2007, 2008, and 2009). For the dominant species (Fundulus diaphanus, Lepomis macrochirus, Etheostoma olmstedi, Morone americana, Lepomis gibbosus, and Fundulus heteroclitus), density was nearly always higher in SAV, but overall, species richness was highest in shoreline habitats sampled with seines. Although historical monitoring of fish in Gunston Cove (and throughout Chesapeake Bay) is based upon seine sampling (and trawl sampling in deeper areas), the high densities of fish and larger areal extent of SAV indicated that complementary sampling of SAV habitats would produce more accurate trends in abundances of common species. Because drop ring samples cover much less area than seines and may miss rare species, a combination of methods that includes seine sampling is needed for biodiversity assessment. The resurgence of SAV in tidal freshwater signifies improving water quality, and methods we evaluated here support improved inferences about population trends and fish community structure as indicators of ecosystem condition.     

Monitoring fish communities in wadeable lowland streams: comparing the efficiency of electrofishing methods at contrasting fish assemblages

Electrofishing is considered a reliable tool to assess the assemblages and biodiversity of fish in wadeable streams. The most widely used electrofishing techniques (point [P], single-pass [S-P], and multiple-pass [M-P]) vary as to the effort needed for sample collection, and this may potentially influence the degree of accuracy. Moreover, little is known about the comparability of the methods and their specific performance in streams with different fish assemblages. The aim of this investigation was to validate (using M-P sampling as reference) the use of P and S-P electrofishing techniques to accurately assess the richness, density and size distribution of fishes in small streams at both regional and global scale independently of fish assemblages and geographical region. We sampled 50-m-long reaches in a total of 33 lowland stream reaches that were located in different climatic and biogeographical regions (Uruguay and Denmark) and hosted different fish assemblages. Subtropical fish communities exhibited higher richness (Uy: 12–32, Dk: 1–9) and densities (Uy: 1.3–5.2, DK: 0.1–4.9 in. m^?2) than temperate streams. We applied both "global models" using the entire database (33 sites) and "local models" including the same number of sites but using the climatic region as a model variable. Regression analyses revealed that the P, S-P and M-P methods all provided an adequate picture of the species composition and size distribution, and transfer equations for comparison between methods are thus not required. Conversely, richness was better predicted by S-P and by P techniques for regional and global models, respectively. Transfer equations obtained for abundance revealed that the P and S-P models can accurately transform catch data into M-P estimations. The transfer equations provided here may have great relevance as they allow relatively reliable comparisons to be made between data obtained by different techniques. We also show that less intensive sampling techniques may be equally useful for monitoring purposes as those requiring more intensive efforts (and costs). We encourage validation of our developed transfer equations on data from other regions of the world.     

Can tree species diversity be assessed with Landsat data in a temperate forest?

The diversity of forest trees as an indicator of ecosystem health can be assessed using the spectral characteristics of plant communities through remote sensing data. The objectives of this study were to investigate alpha and beta tree diversity using Landsat data for six dates in the Gönen dam watershed of Turkey. We used richness and the Shannon and Simpson diversity indices to calculate tree alpha diversity. We also represented the relationship between beta diversity and remotely sensed data using species composition similarity and spectral distance similarity of sampling plots via quantile regression. A total of 99 sampling units, each 20 m × 20 m, were selected using geographically stratified random sampling method. Within each plot, the tree species were identified, and all of the trees with a diameter at breast height (dbh) larger than 7 cm were measured. Presence/absence and abundance data (tree species number and tree species basal area) of tree species were used to determine the relationship between richness and the Shannon and Simpson diversity indices, which were computed with ground field data, and spectral variables derived (2 × 2 pixels and 3 × 3 pixels) from Landsat 8 OLI data. The Shannon-Weiner index had the highest correlation. For all six dates, NDVI (normalized difference vegetation index) was the spectral variable most strongly correlated with the Shannon index and the tree diversity variables. The Ratio of green to red (VI) was the spectral variable least correlated with the tree diversity variables and the Shannon basal area. In both beta diversity curves, the slope of the OLS regression was low, while in the upper quantile, it was approximately twice the lower quantiles. The Jaccard index is closed to one with little difference in both two beta diversity approaches. This result is due to increasing the similarity between the sampling plots when they are located close to each other. The intercept differences between two investigated beta diversity were strongly related to the development stage of a number of sampling plots in the tree species basal area method. To obtain beta diversity, the tree basal area method indicates better result than the tree species number method at representing similarity of regions which are located close together. In conclusion, NDVI is helpful for estimating the alpha diversity of trees over large areas when the vegetation is at the maximum growing season. Beta diversity could be obtained with the spectral heterogeneity of Landsat data. Future tree diversity studies using remote sensing data should select data sets when vegetation is at the maximum growing season. Also, forest tree diversity investigations can be identified by using higher-resolution remote sensing data such as ESA Sentinel 2 data which is freely available since June 2015.     

基于AGNES算法优化BP神经网络和GIS系统的大气污染物浓度预测

建立了大气污染物浓度与影响因子之间的BP神经网络,对城市中各监测点位的次日大气污染物浓度进行预测,采用GIS的插值分析进行污染物空间分布预测,其中BP神经网络的输入向量采用AGNES算法进行处理。以太原市区SO2、PM10浓度预测为例,选择气温、湿度、降水量、大气压强、风速和前5天的污染物浓度等10个参数训练BP神经网络,结果表明,BP神经网络的训练效果较好,预测结果与实际浓度显著相关,R2分别为0.988、0.976;结合太原市8个监测点位的污染物浓度预测值,运用GIS空间差值法绘出SO2、PM10的浓度分布预测图,该图与实际情况大体符合,并且与国控大气污染企业的分布显著相关,Pearson相关系数分别为0.969、0.949。     

山区土壤重金属空间插值性能的改进

山区土壤重金属监测中,密集均匀的布点采样困难大,成本高。根据稀疏非均匀样本数据准确预测山区土壤重金属空间分布,对可视化描述检测元素的分布及趋势具有重要意义。基于云南省楚雄市南部山区表层土壤中重金属锰、钒的采样数据,对集成径向基函数神经网络(IRBFANNs)和传统插值方法:反距离(IDW)、普通克里格(OK)、径向基函数神经网络(RBFANNs),进行了不同等级采样密度下的插值性能比较。结果表明,样本点数量减少时,传统预测方法的插值精度都有所下降,而IRBFANNs算法在样本点较少情况下能够保持最好的插值精确度和稳定性,有效改进了空间插值性能。     

Geographic Analysis of Species Richness and Community Attributes of Forest Birds from Survey Data in the Mid-Atlantic Integrated Assessment Region

Species richness of local communities is a state variable commonly used in community ecology and conservation biology. Investigation of spatial and temporal variations in richness and identification of factors associated with these variations form a basis for specifying management plans, evaluating these plans, and for testing hypotheses of theoretical interest. However, estimation of species richness is not trivial: species can be missed by investigators during sampling sessions. Sampling artifacts can lead to erroneous conclusions on spatial and temporal variation in species richness. Here we use data from the North American Breeding Bird Survey to estimate parameters describing the state of bird communities in the Mid-Atlantic Assessment (MAIA) region: species richness, extinction probability, turnover and relative species richness. We use a recently developed approach to estimation of species richness and related parameters that does not require the assumption that all the species are detected during sampling efforts. The information presented here is intended to visualize the state of bird communities in the MAIA region. We provide information on 1975 and 1990. We also quantified the changes between these years. We summarized and mapped the community attributes at a scale of management interest (watershed units).     

Effect of diel period and season on seining effort required to detect changes in Lake Erie beach fish assemblages

Using beach seining data from Lake Erie beaches, we examined the effect of diel period and season on two common study objectives considered sensitive to sampling effort: (1) has fish assemblage composition changed between two sampling time periods?, and (2) how many samples are necessary to provide a representative sample of the fish assemblage? Across a range of effect sizes and power levels, sample sizes were estimated for species richness, total fish abundance, round goby (Neogobius melanostomus) abundance and emerald shiner (Notropis atherinoides) abundance. For most sampling periods and effect sizes, several hundred to several thousand seine hauls were estimated to be required for detecting large-scale (e.g. basin-wide) changes in round goby and emerald shiner abundance. Depending on the indicator of interest, night seining (either spring or fall) was interpreted as the most efficient approach. For both seasons, sampling effort beyond 40 to 50 night seine hauls results in few additional species being detected. Both species accumulation curves and estimated sample sizes indicate that less effort is required to detect species richness changes than fish abundance. For most effect sizes and power levels, estimates of the number of seine hauls required to detect changes in fish abundance are unrealistically large for most monitoring programs. More practical monitoring strategies could be achieved by adopting alternative indicators (e.g. guilds), a more liberal significance level (p=0.1), a paired-site sampling design, or including reference sites.     

Responses of aquatic macrophytes to anthropogenic pressures: comparison between macrophyte metrics and indices

Macrophyte responses to anthropogenic pressures in two rivers of Central Spain were assessed to check if simple metrics can exhibit a greater discriminatory and explanatory power than complex indices at small spatial scales. Field surveys were undertaken during the summer of 2014 (Duraton River) and the spring of 2015 (Tajuña River). Aquatic macrophytes were sampled using a sampling square (45?×?45 cm). In the middle Duraton River, macrophytes responded positively to the presence of a hydropower dam and a small weir, with Myriophyllum spicatum and Potamogeton pectinatus being relatively favored. Index of Macrophytes (IM) was better than Macroscopic Aquatic Vegetation Index (MAVI) and Fluvial Macrophyte Index (FMI) in detecting these responses, showing positive and significant correlations with total coverage, species richness, and species diversity. In the upper Tajuña River, macrophytes responded both negatively and positively to the occurrence of a trout farm effluent and a small weir, with Leptodictyum riparium and Veronica anagallis-aquatica being relatively favored. Although IM, MAVI, and FMI detected both negative and positive responses, correlations of IM with total coverage, species richness, and species diversity were higher. Species evenness was not sensitive enough to detect either positive or negative responses of aquatic macrophytes along the study areas. Overall, traditional and simple metrics (species composition, total coverage, species richness, species diversity) exhibited a greater discriminatory and explanatory power than more recent and complex indices (IM, MAVI, FMI) when assessing responses of aquatic macrophytes to anthropogenic pressures at impacted specific sites.     

Effects of Clear-Cutting and Plant Competition Control Methods on Carabid (Coleoptera: Carabidae) Assemblages in Northwestern Ontario

Total captures, species richness, diversity, and assemblages of adult carabids (Coleoptera: Carabidae) in a boreal mixed-wood ecosystem study were compared among unharvested, and clear-cut plots untreated or treated with herbicides (Vision® and Release®), mechanical treatments (brush saw and Silvana Selective). Carabids were sampled using unbaited pitfall traps for one growing season 7 and 9 years after clear-cutting, in the second growing season after competition control was applied. A total of 5032 individuals representing 30 species were collected between June 1 and September 25, 1995. Total catches were not affected by the harvest or the conifer release alternatives applied. Species richness and diversity increased following conifer release alternatives. Ten species showed habitat preferences after treatment whereas the remaining 20 species were not significantly affected by treatment.     

Estimating vertebrate,benthic macroinvertebrate,and diatom taxa richness in raftable Pacific Northwest rivers for bioassessment purposes

The number of sites sampled must be considered when determining the effort necessary for adequately assessing taxa richness in an ecosystem for bioassessment purposes; however, there have been few studies concerning the number of sites necessary for bioassessment of large rivers. We evaluated the effect of sample size (i.e., number of sites) necessary to collect vertebrate (fish and aquatic amphibians), macroinvertebrate, and diatom taxa from seven large rivers in Oregon and Washington, USA during the summers of 2006–2008. We used Monte Carlo simulation to determine the number of sites needed to collect 90–95% of the taxa 75–95% of the time from 20 randomly located sites on each river. The river wetted widths varied from 27.8 to 126.0 m, mean substrate size varied from 1 to 10 cm, and mainstem distances sampled varied from 87 to 254 km. We sampled vertebrates at each site (i.e., 50 times the mean wetted channel width) by nearshore-raft electrofishing. We sampled benthic macroinvertebrates nearshore through the use of a 500-μm mesh kick net at 11 systematic stations. From each site composite sample, we identified a target of 500 macroinvertebrate individuals to the lowest possible taxon, usually genus. We sampled benthic diatoms nearshore at the same 11 stations from a 12-cm² area. At each station, we sucked diatoms from soft substrate into a 60-ml syringe or brushed them off a rock and rinsed them with river water into the same jar. We counted a minimum of 600 valves at 1,000× magnification for each site. We collected 120–211 diatom taxa, 98–128 macroinvertebrate taxa, and 14–33 vertebrate species per river. To collect 90-95% of the taxa 75-95% of the time that were collected at 20 sites, it was necessary to sample 11–16 randomly distributed sites for vertebrates, 13–17 sites for macroinvertebrates, and 16–18 sites for diatoms. We conclude that 12–16 randomly distributed sites are needed for cost-efficient sampling of vertebrate richness in the main stems of our study rivers, but 20 sites markedly underestimates the species richness of benthic macroinvertebrates and diatoms in those rivers.     

Assessing changes in the Presque Isle Bay watershed fish community using a modified index of biotic integrity: before and after the elimination of combined sewer overflows

An index of biotic integrity and species richness were used to assess changes in the Presque Isle Bay watershed fish community before and after the elimination of combined sewer overflows (CSOs). The fish community was sampled with a backpack electrofisher in 2011 at 12 stream locations on 4 tributaries of Presque Isle Bay, Erie County, Pennsylvania. All sites were previously sampled in 2001. Significant increases in species richness and index of biotic integrity (IBI) scores were observed in 2011 compared to 2001. The significant increases in species richness and IBI scores occurred following the elimination of 10 CSOs to Garrison Run, 7 CSOs to Cascade Creek, and 37 CSOs to Mill Creek. Despite the increased richness and IBI scores, the fish community remains in poor condition, which may be related to the highly urbanized land use of the watershed. Urban land uses comprise 77 % of the Presque Isle Bay watershed, and in both 2011 and 2001, the watershed as a whole did not meet warm-water habitat criteria. It is unlikely that the fish community will continue to recover without addressing urbanization throughout the watershed.     

Seasonal and Interannual Variations in the Frequency and Severity of Deformities in Larvae of Chironomus riparius (Meigen, 1804) and Prodiamesa olivacea (Meigen, 1818) (Diptera, Chironomidae) collected in a polluted site

Larvae of Chironomus riparius and Prodiamesa olivacea were sampled at intervals overthe period November 1993 and October 1995 from theriver Sar in Galicia (N.W. Spain), with the aim ofinvestigating possible temporal variations in thefrequency of deformed individuals and in `Toxic Score'values (Lenat, 1993), a score for severity of theabnormalities calculated for each species. The resultsobtained indicate that, in both species, deformitiesare more frequent during the colder period of theyear. Similarly, Toxic Score values are higher duringthis period. These finding suggest that studies usingchironomid larvae as indicators of contaminationsshould involve sampling over an extended period, tominimize the influence of temporal variation.     

Use of PFU protozoan community structural and functional characteristics in assessment of water quality in a large, highly polluted freshwater lake in China

Structural and functional parameters of protozoan communities colonizing on PFU (polyurethane foam unit) artificial substrate were assessed as indicators of water quality in the Chaohu Lake, a large, shallow and highly polluted freshwater lake in China. Protozoan communities were sampled 1, 3, 6, 9 and 14 days after exposure of PFU artificial substrate in the lake during October 2003. Four study stations with the different water quality gradient changes along the lake were distinguishable in terms of differences in the community's structural (species richness, individual abundance, etc.) and functional parameters (protozoan colonization rates on PFU). The concentrations of TP, TN, COD and BOD as the main chemical indicators of pollution at the four sampling sites were also obtained each year during 2002-2003 for comparison with biological parameters. The results showed that the species richness and PFU colonization rate decreased as pollution intensity increased and that the Margalef diversity index values calculated at four sampling sites also related to water quality. The three functional parameters based on the PFU colonization process, that is, S(eq), G and T90%, were strongly related to the pollution status of the water. The number of protozoan species colonizing on PFU after exposure of 1 to 3 days was found to give a clear comparative indication of the water quality at the four sampling stations. The research provides further evidence that the protozoan community may be utilized effectively in the assessment of water quality and that the PFU method furnishes rapid, cost-effective and reliable information that may be useful for measuring responses to pollution stress in aquatic ecosystems.     

基于BP神经网络的藻类水华预测模型研究

以宁波大学校内池塘2009年3—10月间30周的监测数据为基础,运用BP人工神经网络方法构建预测模型,探求颤藻生物量与总氮、总磷、透明度等6项环境因子之间的关系,选出最佳预测模型,并对模型进行敏感度分析。结果显示:①BP神经网络模型对颤藻生物量预测值与实测值之间拟合程度良好,相关系数达到了0.984,说明BP神经网络模型可以用于水体中藻类水华的短期预测。②通过对构建的BP神经网络模型进行敏感度分析,阐明了宁波大学校内池塘藻类水华的主要驱动因素,并指出控制水体的pH是宁波大学校内池塘藻类水华防治工作的重点。