An indicator system to assess ecological integrity of managed forests

Ecological integrity of managed forests includes the ability of an ecosystem to support a community of organisms with a similar species composition and functional organization as found in nearby natural systems. We developed an indicator system for ecological integrity based on simulated natural disturbance and indicator species to test if forest condition and habitat in managed forests are similar to that found or expected in natural systems. We then applied the method in an area of the boreal forest (Ontario, Canada) where the objective of Ontario's strategic forest management planning approach is, in part, to conserve ecological integrity through the emulation of the natural disturbance process. Forest condition controls the supply of habitat to support the diversity of native organisms, and historically in boreal forests the natural disturbance process drove forest condition. We selected indicators of forest condition (landscape pattern and compositional mosaic) and habitat function (occupancy rates for a broad range of forest birds), and applied our assessment system to test whether indicators of forest condition and habitat function reflect outcomes expected if natural disturbance processes were successfully emulated. We collected occupancy data in natural and managed forest disturbance types using autonomous acoustic recorders, applied occupancy/detection modeling to estimate corrected occupancy rates (ψ), and then tested for differences in ψ between disturbance types. Some indicators of forest condition were within the range expected under natural disturbance, but we found relatively less old conifer, more young deciduous and greater edge density in managed forests relative to forests of natural disturbance origin. Most species (11 of 14) occurred with equal ψ in habitat originating from the two disturbance types. Brown creeper (Certhia americana), bay-breasted warbler (Mniotilta varia) and red-eyed vireo (Vireo olivaceus) differed between disturbance types. Brown creeper uses older conifer and occurred at lower rates in managed forest, while red-eyed vireo uses a range of deciduous forest ages, and occurred at higher rates in managed forest. Differences in quantity and/or quality of specific habitat types likely explain the responses. The results suggest what directional changes in the forest pattern and compositional mosaic would improve ecological similarity with natural systems, but also indicate what further research is required. We believe this approach to assessing ecological integrity can be adapted to study the effectiveness of conservation management strategies in other systems, and will contribute to adaptive management approaches and evidence-based policy development.     

Fungal tolerance to Congo red,a cell wall integrity stress,as a promising indicator of ecological niche

Differential sensitivities to the cell wall stress caused by Congo red (CR) have been observed in many fungal species. In this study, the tolerances and sensitivities to CR was studied with an assorted collection of fungal species from three phylogenetic classes: Sordariomycetes, Dothideomycetes, and Eurotiomycetes, three orders, and eight families. These grouped into different ecological niches, such as insect pathogens, plant pathogens, saprotrophs, and mycoparasitics. The saprotroph Aspergillus niger and the mycoparasite Trichoderma atroviride stood out as the most resistant species to cell wall stress caused by CR, followed by the plant pathogenic fungi, a mycoparasite, and other saprotrophs. The insect pathogens had low tolerance to CR. The insect pathogens Metarhizium acridum and Cordyceps fumosorosea were the most sensitive to CR. In conclusion, Congo red tolerance may reflect ecological niche, accordingly, the tolerances of the fungal species to Congo red were closely aligned with their ecology.     

Catchment land cover as a proxy for macroinvertebrate assemblage structure in Carpathian Mountain streams

We compared land cover, riparian vegetation, and instream habitat characteristics with stream macroinvertebrate assemblages in 25 catchments in the Carpathian Mountains in Central Europe. This study area was particularly selected because of its diverse history of forest and agricultural ecosystems linked to geopolitical dynamic, which provide a suite of unique landscape scale, land cover settings in one ecoregion. Canonical Correspondence Analysis (CCA) showed that variation in composition and structure of macroinvertebrate assemblages was primarily related to four land cover types, and not to riparian or instream habitat. These were the portions in the catchment areas of (1) broadleaved forest, (2) fine-grained agricultural landscape mosaic with scattered trees (e.g., pre-industrial cultural landscape), (3) mixed forest, and (4) natural grassland without trees. Principal Component Analysis (PCA) suggested that land cover types and stream channel substrates co-varied. The PCA also showed that chemical variables, including organic carbon, had higher values in the agricultural landscape compared to natural forests. The major source of variation among taxa in streams was higher abundance of Diptera in agricultural landscapes and of Plecoptera, Coleoptera, Trichoptera, and Amphipoda in forests. Gastropoda and Oligochaeta were more abundant in open, fine-grained agricultural landscape mosaics with scattered trees. Ephemeroptera taxa were quite indifferent to these gradients in catchment land cover, but showed a tendency of being more abundant in the pre-industrial cultural landscape. Our findings suggest that land cover can be used as a proxy of the composition and structure of macroinvertebrate assemblages. This means that land use management at the catchment scale is needed for efficient conservation and recovery of stream invertebrate communities.     

Forest classification in an Amazonian rainforest landscape using pteridophytes as indicator species

Habitat classification systems are poorly developed for tropical rainforests, where extremely high plant species richness causes numerous methodological difficulties. We used an indicator species approach to classify primary rainforest vegetation for purposes of comparative wildlife habitat studies. We documented species composition of pteridophytes (ferns and fern allies) in 635 plots (2×100 m) along 8 transects within a continuous rainforest landscape in northeastern Peruvian Amazonia. Considerable floristic variation was found when the data were analyzed using multivariate methods. The obtained forest classification was interpreted with the help of indicator value analysis and known soil preferences of the pteridophyte species. The final classification included four forest types: 1) inundated forests, 2) terrace forests, 3) intermediate tierra firme forests and 4) Pebas Formation forests. This rapid and relatively simple vegetation classification technique offers a practical, quantitative method for large-scale vegetation inventory in complex rainforest landscapes.     

Using macroinvertebrates to assess ecological integrity of streams remediated for acid mine drainage

Studies assessing stream restoration of acid mine drainage (AMD) usually use community metrics that emphasize only a few attributes of ecological integrity. We used a variety of measures to examine recovery of macroinvertebrate communities in a watershed receiving passive treatment for AMD. We measured density, richness, functional diversity (FD), taxonomic diversity (Δ, a measure of taxonomic relatedness), and community taxonomic composition at 4 stream sites over an 11‐year period: 1 reference not affected by AMD, 2 downstream of treatment systems, and 1 AMD‐impacted site with no treatment. Our objectives were to determine if communities have responded to remediation, and examine whether FD and Δ provided additional insights into ecology integrity. Richness, FD, and Δ at the reference site were significantly higher than for other sites, but did not consistently differentiate treated and untreated sites. Additional information provided by FD and Δ on attributes related to functional redundancy and taxonomic relatedness was self‐evident given the simplicity of nonreference communities. Ordination of community composition incorporating knowledge of taxa autecologies indicated that treated sites were slightly less impacted than the untreated site, and trend analysis indicated slight improvement at 1 treated site. However, communities below treatment remained taxonomically distinct, with much lower densities than reference condition. Macroinvertebrate communities showed less recovery than was determined for benthic diatom communities at the same sites in a previous study. Several attributes of ecological integrity for different types of stream organisms should be incorporated into bioassessment of AMD restoration because they can vary in their response.     

Conditional heteroscedasticity as a leading indicator of ecological regime shifts

Regime shifts are massive, often irreversible, rearrangements of nonlinear ecological processes that occur when systems pass critical transition points. Ecological regime shifts sometimes have severe consequences for human well-being, including eutrophication in lakes, desertification, and species extinctions. Theoretical and laboratory evidence suggests that statistical anomalies may be detectable leading indicators of regime shifts in ecological time series, making it possible to foresee and potentially avert incipient regime shifts. Conditional heteroscedasticity is persistent variance characteristic of time series with clustered volatility. Here, we analyze conditional heteroscedasticity as a potential leading indicator of regime shifts in ecological time series. We evaluate conditional heteroscedasticity by using ecological models with and without four types of critical transition. On approaching transition points, all time series contain significant conditional heteroscedasticity. This signal is detected hundreds of time steps in advance of the regime shift. Time series without regime shifts do not have significant conditional heteroscedasticity. Because probability values are easily associated with tests for conditional heteroscedasticity, detection of false positives in time series without regime shifts is minimized. This property reduces the need for a reference system to compare with the perturbed system.     

Non-native fish species are related to the loss of ecological integrity in Neotropical streams: a multimetric approach

Hydrobiologia - The use of multimetric indices (MMIs) became a common practice for evaluating ecological integrity of aquatic ecosystems. However, most developed MMIs ignore non-native species as...     

Forest reconstruction as ecological engineering

Land restoration involves reconstruction of the native biota in a sustainable form. If reconstruction involves deliberate manipulation of biological organisms and the physical-chemical environment to achieve specific human goals, it qualifies as ecological engineering. Restoration which uses natural processes to achieve endpoints which are unpredictable but can be accepted because they are “natural” is not ecological engineering. In Japan a system of forest reconstruction has been developed which is based on knowledge of the potential vegetation of a site, knowledge of the methods of germination and growth of the species which compose the mature vegetation and a method of site preparation and planting. This ecological engineering approach has been used on 285 sites, in a variety of habitats, to form dense stands of vegetation to hide industrial complexes, control visual, noise and chemical pollution, stabilize soil and beaches and provide urban green space. The technique has also been used to restore tropical rain forest.     

A new biological indicator to assess the ecological status of Mediterranean trout type streams

Fuelled by the generalized degradation of freshwater ecosystems, the development of tools to assess their ecological status has been the focus of intensive research in the last decades. Although fish are one of the key biological quality elements used to describe the ecological status of rivers, fish metrics that accurately respond to disturbances in Mediterranean trout type streams are still lacking. In these systems, multimetric indices are not optimal indicators because of their low species richness and abundances, thus alternative approaches are needed. Since carrying capacity defines the potential maximum abundance of fish that can be sustained by a river, its relationship with actual density (D/K ratio) could be an accurate indicator of population conservation status and consequently of the ecological status of the river. Based on this rationale, we modeled carrying capacity dynamics for 37 brown trout populations during a 12-year study period. We analyzed the response of the D/K ratio to a gradient of increasing environmental harshness and degradation in order to assess its suitability to accurately measure brown trout conservation status. Our results showed that the D/K ratio was highly sensitive to temporal and spatial variations in environmental conditions and the levels of human-induced environmental degradation. Variations in the environmental and human degradation factors included in the best explaining regression models developed for the whole population and by age classes accounted for between 58 and 81% of the variation in the D/K ratio. Likewise, the D/K ratio was sensitive to both general and life stage specific disturbance factors. Further analyses helped identify the factors limiting population abundance. Therefore, the D/K ratio could be an interesting indicator to consider when defining objective management plans and corrective actions in degraded rivers and streams subject to Mediterranean climatic conditions.     

The utility of DNA barcoding as a tool to assess the success of ecological restoration using Hemiptera as a biological indicator

DNA barcoding uses a short, standardized DNA fragment to sort individuals into species. This molecular technique has applications in fields including ecology, evolution, conservation, and biogeography. In ecological applications such as species monitoring and habitat restoration, its potential has not been fully realized and implemented. Invertebrates are excellent biological indicators, as changes in species diversity or community assemblage provide important insights into the condition of, or changes in, the environment. This information is particularly useful within the context of restoration ecology. In this study, DNA barcoding is used to assess the potential of Hemiptera as a biological indicator of restoration success for the Buffelsdraai Landfill Site Community Reforestation Project (Durban, South Africa). A total of 393 Hemiptera specimens were collected from sites reforested at distinct phases (plots reforested in 2010, 2012, and 2015) and two reference sites (natural forest and grassland). The Hemiptera species composition and assemblage were assessed by analyzing diversity indices, ordination, unweighted pair‐group average cluster analysis, and phylogenetic analysis. Hemiptera species composition varied significantly across the chronologically different reforested sites, with a higher species richness observed in the older reforested plots. This suggests that Hemiptera diversity can be used to track restoration success, even over the small temporal scale used in this study. This study highlights the utility of DNA barcoding as a taxonomic sorting tool both to monitor ecological restoration and to discover specific taxa within Hemiptera that may be useful biological indicators.     

Assessment of streams of the eastern United States using a periphyton index of biotic integrity

Benthic algae were collected from 272 eastern United States streams and rivers and analyzed for diatom species richness and dominance, the relative abundance of acidobiontic, eutraphentic, and motile diatoms, standing crops of chlorophyll and biomass, and alkaline phosphatase activity. These data were used to calculate a periphyton index of biotic integrity (PIBI), and values of the index were compared among reference, moderately impacted, and disturbed streams. The level of disturbance was based on stream chemistry, riparian disturbance, or a combined classification. Analyses of variance showed that PIBI was significantly higher in reference streams for all classifications. The PIBI and its metrics were correlated with many of the chemistry and habitat variables, and canonical correlation analysis revealed three significant environmental gradients which extracted 84% of the variance in the PIBI and its metrics. We used the mean 75th, 25th, and 5th percentile scores from the reference sites to set thresholds for excellent, good, fair, or poor condition. Applying these criteria to the cumulative distribution of total stream length in the region, we found that 4.3% of the stream length was in excellent condition; 20.8% in good condition; 56.4% in fair condition; and 18.5% in poor condition. The sensitivity of the PIBI and its component metrics to environmental stressors supports the use of this index for monitoring ecological conditions in streams in the eastern United States and as a tool to aid in diagnosing the causes of their impairment.     

Seagrass reproductive effort as an ecological indicator of disturbance

The available information on the changes in the reproductive effort (RE) of seagrasses in response to disturbances was reviewed and analysed to assess if seagrasses invest in RE when disturbed, and if this response is related to specific types of disturbance or seagrass traits. In 72% of the documented cases RE increased with disturbance, in 25% it decreased, and in 3% no changes were reported. Overall, seagrass RE increased 4-fold with disturbance. Anthropogenic disturbances had the highest impact on RE (a 13-fold increase); 3 times higher than the effect of natural disturbances. Mechanical and sedimentary/hydrodynamics disturbances caused the highest RE increase (9- and 5-fold, respectively). The positive RE response was significantly correlated with rhizome diameter of seagrasses, but not with shoot size (mass or length), suggesting that species with higher storage capacity have a higher capacity of investing in sexual reproduction when conditions deteriorate. Seagrasses showed a general trend of increasing RE under disturbance; this was evident regardless of the origin and type of disturbance, which suggests that changes in seagrass RE provide a valuable indicator of disturbance in coastal areas.     

White Fringed Orchid as indicator of Sphagnum bog integrity

Ecosystem integrity often needs to be evaluated rapidly; indicator species may be used to do so. In this study we aim to evaluate whether White Fringed Orchid is an indicator of ombrotrophic peatland (Sphagnum bog) integrity in northeastern North America. The study was carried out in 113 sampling plots located in 60 bogs of southern Quebec, Canada. Vegetation (species cover and strata) and abiotic variables were collected within the plots and disturbance variables were determined using aerial photographs and satellite imagery. The level of integrity on each plot was assessed by calculating a Floristic Quality Assessment Index (FQAI) and an Anthropogenic Activity Index (AAI). Data were analyzed with ANOVAs, multivariate techniques and logistic multiple regression. Results indicated that the presence and abundance of White Fringed Orchid were positively linked to the FQAI and negatively to the perimeter/area ratio variable, suggesting that this species is most abundant and frequent at sites that are less disturbed and influenced by human activities. ANOVA showed that the AAI was lower in plots with high orchid abundance than in those without, suggesting that the orchid grows in more pristine sites. Our study showed that White Fringed Orchid abundance and presence may be used to rapidly and accurately evaluate Sphagnum bog integrity.     

Presence of indicator plant species as a predictor of wetland vegetation integrity: a statistical approach

We fit regression and classification tree models to vegetation data collected from Ohio (USA) wetlands to determine (1) which species best predict Ohio vegetation index of biotic integrity (OVIBI) score and (2) which species best predict high-quality wetlands (OVIBI score >75). The simplest regression tree model predicted OVIBI score based on the occurrence of three plant species: skunk-cabbage (Symplocarpus foetidus), cinnamon fern (Osmunda cinnamomea), and swamp rose (Rosa palustris). The lowest OVIBI scores were best predicted by the absence of the selected plant species rather than by the presence of other species. The simplest classification tree model predicted high-quality wetlands based on the occurrence of two plant species: skunk-cabbage and marsh-fern (Thelypteris palustris). The overall misclassification rate from this tree was 13 %. Again, low-quality wetlands were better predicted than high-quality wetlands by the absence of selected species rather than the presence of other species using the classification tree model. Our results suggest that a species’ wetland status classification and coefficient of conservatism are of little use in predicting wetland quality. A simple, statistically derived species checklist such as the one created in this study could be used by field biologists to quickly and efficiently identify wetland sites likely to be regulated as high-quality, and requiring more intensive field assessments. Alternatively, it can be used for advanced determinations of low-quality wetlands. Agencies can save considerable money by screening wetlands for the presence/absence of such “indicator” species before issuing permits.     

The meiofauna community of two small German streams as indicator of pollution

In order to validate the results of biomarkerstudies as tools for the assessment of smallstreams, the meiofauna community and inparticular the nematodes of two small Germanstreams (Baden-Württemberg) were analysedduring 1998. One site of Krähenbach (AB)and three sites of Körsch (KA, KD, KE) weresampled in March, May, July and November. Thelimnological analysis indicates that two sitesof Körsch (KD, KA: polluted sites) weredistinctly enriched by inorganic and organicsubstances compared with KE and AB (unpollutedsites). In the Krähenbach, nematodes,chironomids, oligochaetes and tardigradesdominated the meiofauna community (92%),whereas in the Körsch nematodes andoligochaetes were dominant (93%). The totalmeiofauna density showed seasonal patterns inboth streams. Nematodes were the most abundantgroup at the four investigated locations. Themean abundance of nematodes ranged from 28individuals per 10 cm² to 1205 individualsper 10 cm² and at least 103 species wereidentified, belonging to 25 families. Nosubstantial difference occurred in thedistribution of nematode feeding-types at thepolluted and unpolluted sites. All communitieswere clearly dominated by bacteria-feedingnematodes with percentages ranging from 70% to75%. In order to assess the sediment quality,the nematode community of each sampling sitewas interpreted using the ratio ofSecernentea/Adenophorea as well as the MaturityIndex (MI). Both the percentage of Adenophoreaand the MI decreased towards the pollutedsites, detecting a more disturbed nematodecommunity at the polluted than unpollutedsites. However, there are nematode species(e.g. Tobrilus diversipapillatus) whichdo not fit in this assessment concept. In thisstudy the disturbance of the nematode communitywas related to eutrophication, rather than tothe effect of contaminants. In summary, thisinvestigation confirmed that meiofaunaorganisms and in particular nematodes aresuitable for the assessment of the sedimentquality of running waters.     

Assessing greenhouse gas emissions from peatlands using vegetation as a proxy

Drained peatlands in temperate Europe are a globally important source of greenhouse gas (GHG) emissions. This article outlines a methodology to assess emissions and emission reductions from peatland rewetting projects using vegetation as a proxy. Vegetation seems well qualified for indicating GHG fluxes from peat soils as it reflects long-term water level, affects GHG emissions via assimilate supply and aerenchyma and allows fine-scaled mapping. The methodology includes mapping of vegetation types characterised by the presence and absence of species groups indicative for specific water level classes. GHG flux values are assigned to the vegetation types following a standardized protocol and using published emission values from plots with similar vegetation and water level in regions with similar climate and flora. Carbon sequestration in trees is accounted for by estimating the annual sequestration in tree biomass from forest inventory data. The method follows the criteria of the Voluntary Carbon Standard and is illustrated using the example of two Belarusian peatlands.     

Use of dietary diversity score as a proxy indicator of nutrient adequacy of rural elderly people in Sri Lanka

ABSTRACT: BACKGROUND: Macro and micro nutrient deficiencies are public health concerns in most developing countries including Sri Lanka, partly due to monotonous, cereal-based diet that lacks diversity. The objective of the study was to assess validity of food variety score (FVS), dietary diversity score (DDS) and dietary serving score (DSS) as indicators of nutrient adequacy of rural elderly people in Sri Lanka. FINDINGS: A sample of 200 apparently healthy elderly people >60y of age were studied. A single 24 h recall was performed to compute dietary diversity indicators. Pearson's correlation was used to assess the utility of FVS, DDS and DSS as indicators of nutrient adequacy. Sensitivity (Se) and specificity (Spe) analysis were done to determine the most appropriate cut-off points for using FVS and DDS to categorize elderly people with adequate nutrient intake. The average (standard deviation) of the food variety score, dietary diversity score and dietary serving score was 8.4 (2), 4.4 (0.9) and 11.4 (2.5), respectively. Mean adequacy ratio (MAR) of 12 nutrients was 0.39 (39 %). Pearson's correlation coefficients between MAR and FVS was 0.45 (P < 0.01), for DDS it was 0.48 (P < 0.01) and for DSS it was 0.58 (P < 0.01). When maximizing sensitivity and specificity, the best cut-off point for achieving 50 % of MAR was about 9 and 4.5 for FVS and DDS, respectively. CONCLUSION: In conclusion, FVS, DDS and DSS were useful proxy indicators of nutrient adequacy of rural elderly people in Sri Lanka. Indeed, the performance of the indicators is improved when considering the quantities of food consumed.     

Characterization of benthic habitat settings in a lagoonal ecosystem using free-living nematodes as proxy

Free-living nematodes are sensitive to most of the disturbances and therefore have ability to reflect direct structural and functional changes in an ecosystem. We studied nematode assemblages of Chilika Lagoon, the largest lagoon of Asia, across spatio-temporal scales in link with environmental variables and evaluated nematode assemblages as a proxy to characterize lagoonal benthic habitat settings. Our results revealed that nematode communities showed significant variation spatially and temporally in terms of mean density (16–854/10 cm²) and mean number of species (7–74). Salinity is the key factor that controls nematode community structure across this lagoon and was strongly supported by statistical analyses. The observed nematode assemblages were further used as a proxy to assign benthic habitats of Chilika into distinct biological, topographical and hydrological settings. This study showed that nematode assemblages could be effectively used for long term ecological monitoring of dynamic sedimentary environment of lagoons globally.     

Papillar integrity as an indicator of stigmatic receptivity in kiwifruit (Actinidia deliciosa)

Stigmatic receptivity is a major factor limiting fruit set inkiwifruit (Actinidia deliciosa). The aim of this work was toknow what determines the cessation of stigmatic receptivityin this species. Stigmatic receptivity has been analysed inkiwifruit through the capacity of the stigma to sustain pollengermination and has been related to stigmatic development anddegeneration.The stigma of kiwifruit flowers has a papillatesurface which from anthesis is covered with an abundant exudate.Papillae are unicellular and contain a number of phenolic deposits.During the lifetime of the flower these papillae gradually loseturgidity, while stigmatic secretion increases until 5 d afteranthesis when papillae start rupturing and the papillar contentis liberated into the germination medium. Stigmatic receptivityis high at anthesis and lasts for the next 4 d. However, itdrastically decreases 5 d after anthesis and it is nil 2 d later.The pattern of stigmatic receptivity closely fits that of papillarintegrity, indicating that stigmatic receptivity relies on thisintegrity. Since papillar integrity can easily be evaluatedin stigmatic arms, this criterion can be used as a quick methodto estimate stigmatic receptivity. Key words: Actinidia deliciosa, kiwifruit, stigma, receptivity, pollen     

Long-term recovery of macroinvertebrate biota in grossly polluted streams: Re-colonisation as a constraint to ecological quality

To meet targets imposed by the European Water Framework Directive (2000/60/EC) it is vital that measures to improve the status of rivers are both effective and economically viable. Achievement of such aims needs robust understanding of biological responses to changes in water quality vis-à-vis mechanisms of and constraints to the colonization of previously polluted sites. This study therefore examined the long-term chemical and biological changes in historically polluted rivers to elucidate the responses of macroinvertebrate biota to improvements in chemical water quality. For three historically polluted sites in the English Midlands, data from surveys over a period of ca. 50 years were analysed. Ammonia (NH₃) and 5-day biochemical oxygen demand (BOD₅) were used as chemical water quality indicators. Variations in the ecological recovery of the study sites were assessed using an average pollution sensitivity score (Average Score Per Taxon) and the number of taxa present (usually to family level) present in hand-net samples. Ecological recovery varied widely and was influenced by the intensity and spatial extent of the pollution and the proximity of available sources of potential colonisers. At the site most isolated from potential sources of colonizing taxa, no clean-water macroinvertebrate taxa were recorded 30 years after the major sources of pollution ceased. Where clean-water colonisers were more readily available, significant improvements in ecological quality followed within 2–5 years of the improvements in chemical quality. Macroinvertebrate communities and hence monitoring data may thus be indicative of long past conditions or of biological isolation rather than contemporaneous chemical conditions. Combined chemical and biological data were used to explore a generic model for predicting recovery rates and success. Neither BOD₅ nor NH₃ were found to provide a consistent and meaningful prediction of either average pollution tolerance of macroinvertebrate taxa or of the number of taxa present. Long-term relationships between macroinvertebrate variables and chemical water quality variables, however, were non-linear, suggesting that water quality thresholds may have to be exceeded before biological recovery can occur. Even when chemical water quality has been improved substantially, the apparent ecological status of macroinvertebrate communities may not reflect reduced pollution levels attained until adequate time to allow for re-colonisation (possibly decades) has elapsed.