Very-broad-scale assessment of human impacts on river condition

1. Management of whole rivers and river catchments requires a comprehensive set of information about river condition and use, both existing and historical, and the links between them at regional, state or national scales. This paper outlines a new approach to the assessment of river condition, using a small team was able to assess 210 000 km of rivers across more than 3 million km² of Australia in little more than a year. 2. The approach was driven by a hierarchical model of river function, which assumed that broad‐scale catchment characteristics affect local hydrology, habitat features, water quality and, ultimately, aquatic biota. The model provided the basis for selecting important ecologically relevant features that indices should represent. For each reach of each river we derived a biological index and an environmental index based on measures quantifying catchment and hydrological condition, and habitat and water quality condition. Data came from existing state and national databases, satellite images, site measurements and process models. 3. All indices were calculated as deviation from a reference condition, were range‐standardised and were divided into equivalent bands of condition. Amalgamation of index components and of sub‐indices was determined by consideration of their ecological effects; for example, general degradation might be additive, but toxic effects of one component would override all others. 4. Several internal and external validation methods were employed, with the all‐important validation of the final assessments undertaken by comparison with a similar index based on locally measured data. 5. The environmental assessment classified 14% of reaches as largely unmodified, 67% as moderately modified and 19% as substantially modified by human impacts. The biological assessment based on site assessments and modelled data using invertebrates indicated that 70% of reaches were equivalent to reference condition and that 30% were significantly impaired. Catchment disturbance, elevated sediment and nutrient loads, and habitat degradation all contributed to these results. These impacts have all occurred during the last 200 years (post‐European settlement). 6. Partly as a result of the assessments of this study the Australian Government has begun to adopt a more environmentally sustainable approach to broad‐scale water management.     

Management options for river conservation planning: condition and conservation re-visited

1. Systematic conservation planning is a process widely used in terrestrial and marine environments. A principal goal is to establish a network of protected areas representing the full variety of species or ecosystems. We suggest considering three key attributes of a catchment when planning for aquatic conservation: irreplaceability, condition and vulnerability. 2. Based on observed and modelled distributions of 367 invertebrates in the Australian state of Victoria, conservation value was measured by calculating an irreplaceability coefficient for 1854 subcatchments. Irreplaceability indicates the likelihood of any subcatchment being needed to achieve conservation targets. We estimated it with a bootstrapped heuristic reserve design algorithm, which included upstream–downstream connectivity rules. The selection metric within the algorithm was total summed rarity, corrected for protected area. 3. Condition was estimated using a stressor gradient approach in which two classes of geographical information system Layers were summarised using principal components analysis. The first class was disturbance measures such as nutrient and sediment budgets, salinisation and weed cover. The second class was land use layers, including classes of forestry, agricultural and urban use. The main gradient, explaining 56% of the variation, could be characterised as agricultural disturbance. Seventy‐five per cent of the study area was classified as disturbed. 4. Our definition of vulnerability was the likelihood of a catchment being exposed to a land use that degrades its condition. This was estimated by comparing land capability and current land use. If land was capable of supporting a land use that would have a more degrading effect on a river than its current tenure, it was classified vulnerable (66% of the study area). 79% of catchments contained more then 50% vulnerable land. 5. When integrating the three measures, two major groups of catchments requiring urgent conservation measures were identified. Seven per cent of catchments were highly irreplaceable, highly vulnerable but in degraded condition. These catchments were flagged for restoration. While most highly irreplaceable catchments in good condition were already protected, 2.5% of catchments in this category are on vulnerable land. These are priority areas for assigning river reserves.     

Efficient nonlinear predictive control of a biochemical reactor using neural models

This paper describes the application of artificial neural networks to modelling and control of a continuous fermentor. A computationally efficient nonlinear model predictive control (MPC) algorithm with nonlinear prediction and linearisation (MPC-NPL) which needs solving on-line a quadratic programming problem is developed. It is demonstrated that the algorithm results in closed-loop control performance similar to that obtained in nonlinear MPC, which hinges on full on-line non-convex optimisation. The computational complexity of the MPC-NPL algorithm is shown, control accuracy and robustness are also demonstrated in the case of noisy measurements and disturbances affecting the process.     

AusRivAS: using macroinvertebrates to assess ecological condition of rivers in Western Australia

1. AusRivAS (Australian River Assessment Scheme) models were developed, using macroinvertebrates as indicators, to assess the ecological condition of rivers in Western Australia as part of an Australia-wide program. The models were based on data from 188 minimally disturbed reference sites and are similar to RIVPACS models used in Britain. The major habitats in the rivers (macrophyte, channel) were sampled separately and macroinvertebrates collected were identified to family level. 2. Laboratory sorting of preserved macroinvertebrate samples recovered about 90% of families present when 150 animals were collected, whereas live picking in the field recovered only 76%. 3. Reference sites clustered into five groups on the basis of macroinvertebrate families present. Using seven physical variables, a discriminant function allocated 73% of sites to the correct classification group. A discriminant function based on seven physical and two chemical variables allocated 81% of sites to the correct group. However, when the same reference sites were re-sampled the following year, the nine variable discriminant function misallocated more sites than the seven variable function, owing to annual fluctuations in water chemistry that were not accompanied by changes in fauna. 4. In preliminary testing, the wet season channel model correctly assessed 80% of reference sites as undisturbed in the year subsequent to model building (10% of sites were expected to rate as disturbed because the 10th percentile was used as the threshold for disturbance). Nine sites from an independent data set, all thought to be disturbed, were assessed as such by the model. Results from twenty test sites, chosen because they represented a wide range of ecological condition, were less clear-cut. In its current state the model reliably distinguishes undisturbed and severely disturbed sites. Subtle impacts are either detected inconsistently or do not affect ecological condition.     

A new generation of predictive models: The added value of hybrid models for manufacturing processes of therapeutic proteins

Due to the lack of complete understanding of metabolic networks and reaction pathways, establishing a universal mechanistic model for mammalian cell culture processes remains a challenge. Contrarily, data-driven approaches for modeling these processes lack extrapolation capabilities. Hybrid modeling is a technique that exploits the synergy between the two modeling methods. Although mammalian cell cultures are among the most relevant processes in biotechnology and indeed looks ideal for hybrid modeling, their application has only been proposed but never developed in the literature. This study provides a quantitative assessment of the improvement brought by hybrid models with respect to the state-of-the-art statistical predictive models in the context of therapeutic protein production. This is illustrated using a dataset obtained from a 3.5 L fed-batch experiment. With the goal to robustly define the process design space, hybrid models reveal a superior capability to predict the time evolution of different process variables using only the initial and process conditions in comparison to the statistical models. Hybrid models not only feature more accurate prediction results but also demonstrate better robustness and extrapolation capabilities. For the future application, this study highlights the added value of hybrid modeling for model-based process optimization and design of experiments.     

Small scale application and assessment of a Index of Biotic Integrity for a large boreal river

Modifications and use of Karr’s Index of Biotic Integrity (IBI) for assessing the effects of anthropogenic impacts to aquatic ecosystems have typically occurred using data collected at the macro scale. However, some non-point sources of habitat degradation occur at small scales. One possible source of perturbation to fish habitat in boreal rivers is the application of rip rap shoreline armouring in human use areas. In this study we assess the use of IBI in a small scale setting and discuss the potential impact of rip rap shoreline armouring. We captured small and juvenile fishes weekly during 2002–2003 from 12 sample sites within the littoral zone of a human use area using a modified beach seine. Principal component analysis (PCA) was performed on the data to examine the relationship between species composition and IBI scores. We also performed PCA on the IBI metrics to assess our modifications. The IBI method produced higher scores for armoured sites than for unarmoured sites. We found a strong rank order correlation (Spearman’s ρ > 0.93; p < 0.001) between the modified IBI scores and the first principal component, suggesting that Karr’s original empirically-based IBI is strongly linked to species composition. We found a high degree of redundancy between the metrics of the IBI that validate our modifications. These results suggest that IBI can be a suitable method for assessing non-point impacts from within a small study area. Our results also indicate that fish habitat from rip rap armoured sites within the test area had consistently higher IBI scores that unarmoured sites.     

Classification of macroinvertebrate communities across drainage basins in Victoria, Australia: consequences of sampling on a broad spatial scale for predictive modelling

1. Spatial scale may influence the interpretation of environmental gradients that underlie classification and ordination analyses of lotic macroinvertebrate communities. This could have important consequences for the spatial scale over which predictive models derived from these multivariate analyses can be applied. 2. Macroinvertebrate community data (identified to genus or species) from edge and main-channel habitats were obtained for sites on rivers from 25 of the 29 drainage basins in Victoria. Trends in community similarity were analysed by carrying out separate multivariate analyses on data from the edge habitats (199 sites) and the main-channel habitats (163 sites). 3. Hierarchical classification (UPGMA) showed that the edge data could be placed into 11 site groups and the main-channel data into 12 site groups. 4. Ordination analysis (hybrid multidimensional scaling) showed no sharp disjunctions between site groups in either habitat; overlap was frequent. Correlation of the ordination patterns with environmental variables showed that edge communities varied longitudinally within a drainage basin and from the east to the west of Victoria. These two trends were superimposed on one another to form a single gradient on the ordination. The taxon richness of edge communities was also related to the species richness of macrophytes at a site. Main-channel communities also displayed a longitudinal and a geographic gradient, but these two gradients were uncorrelated on the ordination. 5. Community similarity only weakly reflected geographic proximity in either habitat. A preliminary subdivision of Victoria into a series of biogeographic regions did not match the pattern of distribution of site groups for the edge habitat, illustrating the difficulties of applying to lotic communities a priori regionalizations based on terrestrial features of the landscape. 6. The longitudinal gradients in the two data sets were commonly observed in data gathered at smaller spatial scales in Victoria. The other gradients (geographic, macrophyte), however, were either not consistently repeated or not evident at smaller spatial scales. At small spatial scales (i.e. within a single drainage basin) gradients were related to variables that varied over restricted ranges, e.g. mean particle size of the substratum. 7. Species richness was very variable when plotted against river slope or distance of site from source; both of these are measures of position on the longitudinal gradients. In contrast to suggestions in the literature, species richness did not show a unimodal trend on these gradients, or any other trend. 8. Environmental gradients (apart from longitudinal gradients) that underlie predictive models of macroinvertebrate distribution are reflections of the spatial scale on which the model has been constructed and cannot be extrapolated to different scales. Models must be suited to the spatial scale over which predictions are required.     

Model choice for phylogeographic inference using a large set of models

Model‐based analyses are common in phylogeographic inference because they parameterize processes such as population division, gene flow and expansion that are of interest to biologists. Approximate Bayesian computation is a model‐based approach that can be customized to any empirical system and used to calculate the relative posterior probability of several models, provided that suitable models can be identified for comparison. The question of how to identify suitable models is explored using data from Plethodon idahoensis, a salamander that inhabits the North American inland northwest temperate rainforest. First, we conduct an ABC analysis using five models suggested by previous research, calculate the relative posterior probabilities and find that a simple model of population isolation has the best fit to the data (PP = 0.70). In contrast to this subjective choice of models to include in the analysis, we also specify models in a more objective manner by simulating prior distributions for 143 models that included panmixia, population isolation, change in effective population size, migration and range expansion. We then identify a smaller subset of models for comparison by generating an expectation of the highest posterior probability that a false model is likely to achieve due to chance and calculate the relative posterior probabilities of only those models that exceed this expected level. A model that parameterized divergence with population expansion and gene flow in one direction offered the best fit to the P. idahoensis data (in contrast to an isolation‐only model from the first analysis). Our investigation demonstrates that the determination of which models to include in ABC model choice experiments is a vital component of model‐based phylogeographic analysis.     

基于K-均值算法模型的区域土壤数值化分类及预测制图

根据封丘县土壤发生学特点遴选质地、有机质、土壤颜色、pH值、电导率和土壤发生层厚度等作为土壤属性向量,运用k-均值算法模型,对研究区40个土壤剖面样本实施数值化分类,并依据《中国土壤系统分类检索》确定算法模型输出的5个中心土壤剖面的系统分类归属。基于40个样本土壤剖面剖面与各中心剖面类型之间的分类距离,应用地统计学手段预测研究区空间任意位置的土壤与各中心剖面分类距离,完成研究区土壤数值化连续分类并在空间上实现可视化表达。在此基础上,运用去模糊化手段,“硬化”连续分类边界,获得可与传统土壤制图互为参比的研究区土壤预测制图,并对输出结果进行了土壤发生学解释。本研究表明,土壤数值化分类手段与地统计学随机模型相结合可以实现区域土壤的空间预测,且预测制图比传统土壤制图蕴含更加丰富的信息。     

Hierarchical models facilitate spatial analysis of large data sets: a case study on invasive plant species in the northeastern United States

Many critical ecological issues require the analysis of large spatial point data sets – for example, modelling species distributions, abundance and spread from survey data. But modelling spatial relationships, especially in large point data sets, presents major computational challenges. We use a novel Bayesian hierarchical statistical approach, 'spatial predictive process' modelling, to predict the distribution of a major invasive plant species, Celastrus orbiculatus , in the northeastern USA. The model runs orders of magnitude faster than traditional geostatistical models on a large data set of c . 4000 points, and performs better than generalized linear models, generalized additive models and geographically weighted regression in cross-validation. We also use this approach to model simultaneously the distributions of a set of four major invasive species in a spatially explicit multivariate model. This multispecies analysis demonstrates that some pairs of species exhibit negative residual spatial covariation, suggesting potential competitive interaction or divergent responses to unmeasured factors.     

Modeling spatial survival data using semiparametric frailty models

We propose a new class of semiparametric frailty models for spatially correlated survival data. Specifically, we extend the ordinary frailty models by allowing random effects accommodating spatial correlations to enter into the baseline hazard function multiplicatively. We prove identifiability of the models and give sufficient regularity conditions. We propose drawing inference based on a marginal rank likelihood. No parametric forms of the baseline hazard need to be assumed in this semiparametric approach. Monte Carlo simulations and the Laplace approach are used to tackle the intractable integral in the likelihood function. Different spatial covariance structures are explored in simulations and the proposed methods are applied to the East Boston Asthma Study to detect prognostic factors leading to childhood asthma.     

Predicting diversity versus community composition of aquatic plants at the river scale

We tested the relative importance of physical versus chemical factors in explaining aquatic plant species diversity and community composition within a temperate lowland river. A total of 38 macrophyte species were identified at 33 sites along the 104 km length of the Rideau River, a National Heritage River of Canada. Species richness ranged from 0 to 15 species per site, and Shannon diversity from 0 to 2.98. Macrophyte species richness and Shannon diversity were significantly related to the physical characteristics of sites. For Shannon diversity, 77% of the increase was explained by an increase in sediment organic content and a decrease in water velocity. For species richness, 70% of the increase was explained by the latter factors in addition to an increase in the littoral zone (0–2 m depth contour) width and planktonic chlorophyll concentrations. River water chemistry did not explain any observed variation in either Shannon diversity or species richness in this moderately enriched system. In contrast to species richness, the physical and chemical variables measured failed to explain variation in community composition. Cluster analysis did not reveal any grouping of species into distinct communities. Canonical correlation analysis showed that environmental variables had minimal effect on the distribution of most species, with only floating-leaved species responding to water velocity. We conclude that physical factors can predict species diversity at the within river scale but not the species composition at a given site, underlying the need to preserve the geomorphological diversity of rivers to maintain plant diversity.     

Diel distribution patterns of fishes in a temperate large lowland river

During the day, juveniles and typical inshore species in the lower Oder River, Germany, inhabited the littoral (0·05–1·50 m deep) while potamal species inhabited the mid‐channel (40–100 m from the banks). During the night, two behavioural groups were distinguished: 1) potamal species and large individuals of other species, moved from offshore to the littoral, and 2) nocturnal foraging species, became active and increased in inshore and offshore habitats. Inshore, the most significant differences at night was the increasing density of silver bream Blicca bjoerkna increasing both total biomass and total length of fishes caught. Offshore, the increasing density of whitefin gudgeon Gobio albipinnatus at night was most significant. Additional night electrofishing improved the assessment of the abundance, age and size structure of typical potamal fish species. It should be incorporated in large river fish monitoring programmes.     

Assessing structural and functional ecosystem condition using leaf breakdown: studies on a polluted river

1. Leaf breakdown rates of Alnus glutinosa were determined and the structure of decomposer assemblages associated with leaves were analysed to assess the effect of pollution on the ecological condition of the Ave River (North‐west Portugal). 2. Increase in organic and inorganic nutrients was associated with an increase in density and a decrease in richness of macroinvertebrates, a dramatic decline in the conidial production of aquatic hyphomycetes, but no major change in the richness of aquatic hyphomycetes. 3. Downstream nutrient enrichment was correlated with accelerated leaf breakdown rates. 4. The degree of functional impairment assessed by the ratio of leaf breakdown rates in coarse‐mesh and fine‐mesh bags was in accordance with the gradient of pollution defined by two biotic indices. 5. This study supports the contention that leaf breakdown experiments are a valuable tool to assess the effect of pollution on the ecological condition of rivers.     

Assessment of river quality models: a review

The paper reviews river quality models on the basis of their conceptualization, processes, strengths and limitations. It analyzes advances in basic research and compares river quality models, namely AQUATOX, Branched Lagrangian Transport Model (BLTM), One Dimensional Riverine Hydrodynamic and Water Quality Model (EPD-RIV1), QUAL2Kw, Water Quality Analysis Simulation Program (WASP) and Water Quality for River-Reservoir Systems (WQRRS). All these models are widely used and ‘mechanistic’ in nature except for BLTM which was selected due its vast ‘useage’. In addition, the paper highlights the types of errors which occur during the modelling exercise. The paper also emphasizes on the pivotal role played by water quality models for development and formulation of various river restoration projects worldwide. The present review also suggests broad recommendation for choosing a river quality model.     

Bayesian model selection for spatial capture–recapture models

A vast amount of ecological knowledge generated over the past two decades has hinged upon the ability of model selection methods to discriminate among various ecological hypotheses. The last decade has seen the rise of Bayesian hierarchical models in ecology. Consequently, commonly used tools, such as the AIC, become largely inapplicable and there appears to be no consensus about a particular model selection tool that can be universally applied. We focus on a specific class of competing Bayesian spatial capture–recapture (SCR) models and apply and evaluate some of the recommended Bayesian model selection tools: (1) Bayes Factor—using (a) Gelfand‐Dey and (b) harmonic mean methods, (2) Deviance Information Criterion (DIC), (3) Watanabe‐Akaike's Information Criterion (WAIC) and (4) posterior predictive loss criterion. In all, we evaluate 25 variants of model selection tools in our study. We evaluate these model selection tools from the standpoint of selecting the “true” model and parameter estimation. In all, we generate 120 simulated data sets using the true model and assess the frequency with which the true model is selected and how well the tool estimates N (population size), a parameter of much importance to ecologists. We find that when information content is low in the data, no particular model selection tool can be recommended to help realize, simultaneously, both the goals of model selection and parameter estimation. But, in general (when we consider both the objectives together), we recommend the use of our application of the Bayes Factor (Gelfand‐Dey with MAP approximation) for Bayesian SCR models. Our study highlights the point that although new model selection tools are emerging (e.g., WAIC) in the applied statistics literature, those tools based on sound theory even under approximation may still perform much better.     

The use of AUSRIVAS predictive models to assess the response of lotic macroinvertebrates to dams in south-east Australia

1. Sites immediately below seven dams in Victoria and 12 in southern New South Wales (NSW) were sampled for macroinvertebrates using rapid bioassessment techniques. Specimens were identified to the lowest taxonomic level. The dams had different operating and discharge regimes and thus potentially different downstream effects on this fauna.
2. The AUStralian RIVer Assessment Scheme (AUSRIVAS) predictive models for macroinvertebrate composition were applied to the families and genera observed at Victorian sites and to the families only at the NSW sites. Lists of predicted taxa, their probabilities of occurrence and ratios of observed to expected taxa (O/E scores) were obtained from these models.
3. The mean O/E score for Victoria was 0.46 at the genus level and 0.63 at the family level; for NSW the mean score at the family level was 0.62. There was no significant difference between family level scores for the two states. Thus about 40% of expected families were missing at all sites. Family level scores were highly correlated ( r =0.84–0.86, P  < 0.001) with numbers of families or species at a site.
4. There was no correlation between family O/E scores (or family and species richness) and the degree to which current discharge levels below a dam differ from the natural patterns. Thus increased hydrological deviation did not result in greater biological degradation. In addition, there was no association between these biological characteristics and whether dams released surface or bottom water.
5. All the dams seemed to cause much the same disruption to the fauna. Of the families predicted to have widespread occurrence, 9–12 were found at most sites (tolerant taxa), while a larger number (14–24) were missing (intolerant taxa). The limited recolonisation below dams may well be because of the fact that dams act as barriers to drift, the most prominent route for invertebrate colonists.     

Microbial respiration within a floodplain aquifer of a large gravel-bed river

1. Aerobic respiration, productivity and the carbon turnover rate of microbial biofilms were determined at hyporheic and phreatic sites in the Kalispell Valley alluvial aquifer along a transect extending 3.9 km laterally from the main channel of the Flathead River, a sixth order river in Montana (U.S.A.). The effect of experimentally increasing bioavailable organic carbon (acetate) on the respiration rate of biofilms in this carbon‐poor [dissolved organic carbon (DOC) < 2 mg L^?1] aquifer was also measured. 2. Chambers containing natural substratum were placed in‐situ and allowed to colonise for 20 weeks. After 4, 12 and 20 weeks, they were taken to the laboratory where oxygen flux was measured in a computer‐controlled, flow‐through respirometry system. 3. Respiration ranged from 0.01 to 0.33 mg O₂ dm^?3 h^?1 across sites, with means ranging from 0.10 to 0.17 mg O₂ dm^?3 h^?1. Productivity estimates ranged from 0.18 to 0.32 mg C dm^?3 day^?1 (mean 0.25, SE 0.03). The total organic carbon (TOC) of the microbial biofilms ranged from 18.2 to 29.7 mg C dm^?3. Turnover rate ranged from 3.2 to 5.6 year^?1 with a mean of 4.2 year^?1. 4. At the hyporheic site very close to the river, respiration did not significantly increase when samples were supplemented with labile carbon. Respiration increased with increasing DOC addition at hyporheic sites more distant from the river, suggesting a carbon‐limitation gradient within the hyporheic zone. Microbes at the phreatic site did not respond to increasing DOC addition, suggesting that the phreatic biofilm is adapted to low carbon availability. 5. Comparing the volume of the alluvial aquifer (about 0.7 km³) to that of the river benthic sediments (to 0.25 m depth, which amounts to about 1.6 × 10^?4 km³) within the Flathead Valley, leads to the conclusion that interstitial microbial productivity is orders of magnitude greater than benthic productivity. Alluvial aquifers are often voluminous and microbial production is an enormous component of ecosystem production in rivers such as the Flathead.