Development of an aquatic toxicity index for macroinvertebrates

Mnisi, Lucky Nhlanhla
Journal Title
Journal ISSN
Volume Title
Rapid biomonitoring protocols employing riverine macroinvertebrates in South Africa utilise the South African Scoring System version 5 (SASS5). The SASS5 was developed as part of the then River Health Programme (RHP) [now River Eco-status Monitoring Programme (REMP)]. The SASS5 index is a cost-effective procedure (utilising limited sampling equipment) that enables speedy evaluation of a riverine ecosystem’s health using macroinvertebrates as biological indicators of water quality and ecosystem health. As a result, the SASS5 (including earlier versions) has been widely accepted by water quality practitioners and is increasingly incorporated into Ecological Reserve determinations. However, the SASS is widely criticised for being a ‘red flag’ indicator of water quality and ecosystem health because it has the ability to show only whether a river is polluted (including the extent of pollution) or not, but cannot differentiate between pollutant types (whether chemical or physical). To trace the pollutants responsible for changes in water quality, practitioners are therefore required to conduct chemical-based water quality assessments. Chemical analyses can provide accurate measures of the magnitudes of chemical substances present in the river water but they do not readily translate into threshold limits supportive or protective of ecosystems. In South Africa the water quality threshold limits for aquatic ecosystems are provided by the South African water quality guidelines for aquatic ecosystems (volume 7). These guidelines provide threshold limits for the protection of the entire aquatic ecosystem constituting of fish, macroinvertebrates, microinvertebrates, algae and plants. These guidelines are therefore too broad for defining protection thresholds supportive of specific subcomponents (i.e. macroinvertebrates) of aquatic ecosystems. The Aquatic Toxicity Index (ATI) for macroinvertebrates was therefore developed for providing threshold limits for physical and chemical stressors protective of freshwater macroinvertebrates. The ATI is expected to aid water quality practitioners working in the Olifants River and catchments with similar land-uses in at least three ways. Firstly, in interpreting the magnitudes of physico-chemical water quality stressors by providing varying levels of protection (threshold limits) (i.e.PC99, PC95, PC90 and PC80) specific to freshwater macroinvertebrates. Secondly, the ATI is expected to aid in the compression of large volumes of water quality data into manageable quantities (descriptor words and grading symbols). Lastly, conventional water quality reports are replete with technical terminology and symbols emanating from water chemistry and ecotoxicology. While reporting of this kind is accessible to water resource specialists, it may constitute an obstacle for non-technical stakeholders (with no training or experience in water chemistry) like policy makers, political decision makers and the public. These groups generally have neither the time nor the training to study and understand a traditional, technical review of water quality data. Water quality indices are capable of eliminating technical language incurred in water quality reports; hence, they are viewed as necessary tools in reaching multiple audiences by bridging the gaps between the extremes of water quality monitoring and reporting. The ATI is expected to enhance not only accessibility and comprehensibility in all these instances, but utility in general too. Differently expressed, the ATI is expected to aid as a water quality-reporting tool that will help water quality practitioners and managers in communicating technical water quality data to multiple stakeholders even those without training and experience in water chemistry and ecotoxicology. The development of the ATI for macroinvertebrates was conducted in two phases. First, the derivation of Protection Concentrations (PCs). The PCs were obtained by fitting Species Sensitivity Distribution (SSD) curves on short-term (24-96 hours) median lethal (LC50) data for freshwater macroinvertebrates collected from databases and scholarly publications. Before the estimation of the PCs, the toxicity data had to undergo a preparatory process. This involved the conversion of metal stressors from total metal concentrations to dissolved fractions. Additionally, metal stressors whose toxicity is known to be dependent on water hardness (cadmium, chromium (III), copper, lead, nickel and zinc) were adjusted to reflect their toxicity at six different levels of water hardness using USEPA conversions algorithms. In addition, all ammonia data were converted to reflect the toxicity of ammonia as TAN at pH = 8 and temperature = 25°C. The second phase of the index development involved the allocation of index categories. This was to enable the discrimination of stressors’ magnitudes into classes. The final product is a five-point scale classification system (A to E) based on four PC levels (PC99, PC95, PC90 and PC80) for freshwater macroinvertebrates obtained by fitting Species Sensitivity Distribution (SSD) functions on the toxicity data. On development, the index was validated using water quality data, riverine macroinvertebrates survey data and flow data collected from the Olifants River catchment between 2015 and 2016. Eight study sites were covered, located in the upper and the lower Olifants system. Four of these were located in the Olifants River main stem and four from four tributaries (Klein Olifants, Blyde, Ga-Selati and Letaba rivers). The evaluation of the Olifants system based on the assessment of variable-by-variable indicated that Site S5 and S2 (lower Ga-Selati and Klein Olifants) were the most degraded sites in the study, respectively. In addition, the study indicated gross elevation of sulfate, nitrates, pH and copper. The evaluation of the Olifants system using the composite ATI for macroinvertebrates indicated that the system was generally in good condition. However, the identification of the lowest rating score indicated that temperature difference from reference conditions, sulfate, nitrate, zinc and lead were the main variables limiting the water quality of the Olifants system. In addition, the sensitivity analysis of the index conducted as part of the validation process of the index, indicated that temperature difference from reference conditions, sulfate and nitrate were the most important variables in the computation of the index. Investigations of the relationships between the ATI for macroinvertebrates, SASS5 metrics, MIRAI and measures of flow variability revealed negligible and statistically insignificant associations. These could mainly be attributed to three reasons. Firstly, sampling difficulty, this resulted from high density of filamentous algae and floating aquatic vegetation (posing physical obstructions to sampling) in the river. Such extraneous factors rendered the SASS5 sampling protocol (benthic/kick method), ineffective because of clogging of sampling net and loss of specimens in the sifting of aquatic plants for macroinvertebrates. Secondly, the filamentous algae interfered with the availability and suitability of habitat for aquatic invertebrates. For instance it covered stones biotopes (stones in current) forming a thick layer of algae on the stones thereby exhibiting the characteristics of vegetation biotopes for most sampling sites, a major impediment for the SASS5, a method that is largely dependent biotope availability. Thirdly, the SASS was developed for organic pollutants while the ATI for macroinvertebrates is largely driven by metal stressors. Because of toxicity data shortages for regional freshwater macroinvertebrates, a compromise between data availability and quality was considered. For example, toxicity data drawn from global sources were used as the base for the PCs and the index. Despite these limitations, the protection concentrations (numerical sensitivity values) that form the base of the index were comparable with published water quality benchmarks from literature and water quality jurisdictions. In addition, the index has the ability to summarise and discriminate (stressors in terms of concentrations and magnitudes) large quantities of water quality data to facilitate interpretation of the quality of a water’s ability to support freshwater macroinvertebrates. Keywords: Aquatic Toxicity Index, freshwater macroinvertebrates, water quality criteria, species sensitivity distribution, Olifants River
A thesis submitted in fulfilment of the requirements for the degree of Doctor of Philosophy in the Faculty of Science. School of Animal, Plant and Environmental Sciences of the University of the Witwatersrand. Johannesburg, 18 May 2018.
Mnisi, Lucky Nhlanhla Mduduzi (2018) Development of an aquatic toxicity index for macroinvertebrates,University of the Witwatersrand, Johannesburg, <>