The role of the Usuthu River as refuge for the aquatic biodiversity of the lower Phongolo system

Abstract

Freshwater riverine floodplain systems are one of the most threatened ecosystems. These riverine floodplain systems are also the most productive and biodiverse freshwater systems in the world. Impacts that threaten these systems include altered flow regimes, and various agricultural, domestic, and industrial activities. Furthermore, these impacts may be exacerbated by extreme events such as drought. South Africa’s largest and most biodiverse floodplain system, the lower Phongolo River Floodplain (PRF), consists of two large rivers contributing to the downstream associated floodplain lakes. Both the Usuthu and Phongolo rivers, the two largest rivers of the PRF, play important roles in the structure and functioning of the lower PRF aquatic systems and join at the confluence within the only protected area in the PRF, the Ndumo Game Reserve (NGR). More importantly, of these two rivers, the Phongolo River is being heavily regulated by the Pongolapoort Dam while the Usuthu River is not regulated to the same extent. The pressure on the Phongolo River’s flow regime was increased by the local drought conditions experienced between 2015 and 2017. Due to these local drought conditions, the most recent flooding of the Phongolo River was in 2022, eight years after the previous flood (2014). The Phongolo River is not only influenced by the dam, but large-scale agricultural activities contribute heavily to degrading the water quality of the river. Over the years, studies have indicated the degradation of water quality of the Phongolo River as a result of these anthropogenic activities. In contrast, the dynamics of the Usuthu River and its natural flow regime remain largely unknown together with its present environmental quality and ecology with the NGR. Similar anthropogenic activities occur within the Usuthu – Phongolo rivers catchment such as agricultural, industrial and domestic activities. Between the two rivers, the Phongolo River has been the most studied with regard to how the Pongolapoort Dam and agricultural activities have impacted downstream aquatic systems in terms of water quality, ecology, ecotoxicology and human health risks. In contrast, the impacts surrounding the Usuthu River have only been studied in the Eswatini region and therefore less is known of the system within the only protected area in the PRF. Furthermore, the sections of both rivers within the NGR are both different as agricultural activities are adjacent to the Phongolo River within the NGR while no agricultural activities are located near the Usuthu River. Recent studies have shown the impact of the Pongolapoort Dam together with the supra-seasonal drought on the water quality of the Phongolo River and as such indicated that water quality from the Phongolo River has changed considerably and has worsened since the 1970’s. The most recent study in the Usuthu River indicated the contribution of agricultural and industrial activities to the upper Usuthu River. Therefore, the main goal of the study was to conduct abiotic and biotic assessments in order

to determine the present environmental quality, and aquatic biodiversity of the Usuthu River within the NGR. The influence the natural flow regime has on the downstream floodplain systems was assessed and whether the Usuthu River can act as a refuge for the aquatic biodiversity of the lower Phongolo Floodplain. The abiotic assessment consisted of water and sediment quality of the various aquatic systems within the PRF. Samples were collected in triplicate from five sites during two surveys (High Flow – HF and Low Flow – LF) within the PRF. Both physical and chemical properties were analysed in the water and sediment samples. Chemical concentrations of various metals from this study were much lower than concentrations in aquatic systems within the Kruger National Park (KNP) that are known to be polluted by extensive mining, industrial and agricultural activities. The highest metal concentrations from the present study were Cu, Fe and Zn as a result of agricultural and industrial practices in the upper catchment. Salt concentrations in Lake Nyamithi was higher compared to the other aquatic systems within the PRF, while the Phongolo River had higher Total Dissolved Solids (TDS) concentrations than the Usuthu River. The Usuthu River was higher in nitrate (NO3) and nitrite (NO2) concentrations and phosphates (PO4) concentrations were higher in the Phongolo River. The primary cause for lower nutrient concentrations in the Phongolo River was attributed to the Pongolapoort Dam that traps nutrient-rich sediments and upstream nutrient supply. The aquatic systems differed in Total Organic Carbon content (TOC) as the floodplain lakes were higher in TOC than their respective rivers with the highest TOC reported in Lake Shokwe. Metal concentrations differed between the systems as higher metal concentrations were measured in Lake Shokwe compared to Lake Nyamithi and its outlet, while the Usuthu River had the highest metal concentrations in sediments during the LF survey. The analyses of physical and chemical properties lead to a wealth of data that can sometimes be difficult to interpret. Environmental indices were used to summarise the vast amount of data and as such, the Aquatic Toxicity Index (ATI) and Sediment Quality Index (SeQI) were used to determine the water and sediment quality of the aquatic systems within the PRF. The value from the indices indicates the aquatic systems suitability for use and the degree to which the systems are impacted. The Usuthu River during the HF survey had the highest score and Lake Nyamithi outlet the lowest of all the aquatic ecosystems. In general, the variables that contributed the most to lowering the ATI scores were Zn for the metals and Total Dissolved Solids (TDS), ammonium (NH4) and phosphate (PO4) for the nutrients. The scores for both floodplain rivers and associated floodplain lakes remained suitable (i.e., a score between 60 and 100) for aquatic biodiversity, with the exception of Lake Nyamithi where its score indicated that water quality is more suitable for hardy fish species such as Oreochromis mossambicus

and Clarias gariepinus. Sediments from the majority of the aquatic ecosystems found to be of acceptable quality (i.e., above 60) and it was only Lake Shokwe and Lake Nyamithi outlet that scored below 60. The variables that contributed to lowering the SeQI scores were Cr, Cu and Ni and was attributed to agricultural activities (Cr and Cu) and coal mines (Ni) in the upper catchment. The biotic assessments consisted of various aspects such as macroinvertebrate community structures, food web and dietary analyses, fish migration, mercury (Hg) bioaccumulation and human health risks. Aquatic biota are influenced by anthropogenic stressors and therefore studying aquatic biota provides valuable information as to the degree that freshwater systems are influenced by anthropogenic activities. Macroinvertebrate community structures were assessed for both floodplain rivers and associated lakes while also determining the influence of the natural flooding regime on macroinvertebrate community structures by means of different multivariate analyses. Community structures between the regulated Phongolo River and unregulated Usuthu River were not different, however, the taxa in the Usuthu River were more sensitive towards pollution while in the Phongolo River, the taxa were more pollutant tolerant. Furthermore, the Phongolo River inside the NGR had a higher number of taxa than outside the NGR. In the floodplain lakes of the PRF, the Phongolo River associated floodplain lakes were higher in diversity compared to Lake Shokwe, however, these floodplain lakes shared similar taxa such as the backswimmers (Anisops sp. A), the water scavenger beetle (Berosus sp.) and Oligochaeta. The 2017 and 2018 surveys in Lake Nyamithi differed in structure and diversity as the 2017 survey had higher number of taxa, Pielou’s Evenness Index, Margalef’s Species Richness, Shannon-Wiener Diversity Index and Simpson’s Index than the 2018 survey. The influence from the Usuthu River was demonstrated through the higher macroinvertebrate diversity of Lake Shokwe during the HF survey. Moreover, due to no flooding of the Phongolo River during the study period, any changes in the macroinvertebrate diversity in Lake Nyamithi indicates the influence from the Usuthu River. Most importantly, variation partitioning analysis showed that in terms of water quality, habitat preference and type of system; habitat preference were the main contributing factor in structuring the macroinvertebrate community structures of aquatic systems within the PRF. Stable carbon and nitrogen isotopes analyses are a useful tools in determining the consumer diets, tracing anthropogenic impacts, trophic relationships, constructing food web models and indicating how energy flows through an aquatic system. The aim of this biotic assessment was to determine the food web structures and consumer diets from fish of the two floodplain rivers

and associated floodplain lakes and to determine the effect of the impaired flow regime on these food web structures and consumer diets. Additionally, the aim was to determine whether there was any biological connectivity (using Strontium (Sr) isotope ratios) between the aquatic systems within the PRF. The key findings were similar to the macroinvertebrate community structures of floodplain rivers; i.e. there was no difference in food web structure between the two rivers, however, different food web components were collected. Furthermore, food web components from the Usuthu River were significantly enriched in nitrogen and correspond with the water quality of the system. Although it is well known that the Phongolo River below the Pongolapoort Dam receives nitrogen inputs through organic fertilisers from the downstream agricultural activities, the impoundment further contributed substantially to the lower nitrogen values. Consumer diets from the Usuthu River comprised of an integrated mixture of food sources with aquatic vegetation (38%) and macroinvertebrates (50%) being consumed the most. Conversely, consumers from the Phongolo River consumed mostly C3 plants (37%) while C4 plants were the least consumed (6%). Dissimilar food web structures were determined in the floodplain lakes within the PRF as food web components (particularly the tigerfish (Hydrocynus vittatus), sharptooth catfish (Clarias gariepinus) and the dwarf tigerfish (Brycinus imberi)) from Lake Shokwe were one trophic level higher than the food web components from Lake Nyamithi. Only one fish species, H. vittatus was at the top of the food web in Lake Shokwe whereas in Lake Nyamithi two species (H. vittatus and C. gariepinus) occupied the top of the food web. Isotopic signatures in food web components from Lake Shokwe were significantly different to Lake Nyamithi’s food web components, with the exception of δ15N signatures in macroinvertebrates. The diet of consumers within Lake Shokwe were similar to its associated river and preferred an integrated mixture of food sources as detritus were consumed the most (46%) and macroinvertebrates were the least consumed (<10%). Between two different surveys (2017 vs 2018), Lake Nyamithi was also dissimilar in food web structures. No primary producers could be collected during the 2018 survey as a result of the low water levels during the 2018 LF survey, whereas during the 2017 survey, a large variety of primary producers such as detritus, biofilm, leaf litter, and plant material (Trapaceae, Poaceae, Cyperaceae and Nymphaeceae) were collected. Furthermore, H. vittatus and C. gariepinus occupied the top of the food web during the 2018 survey while in the 2017 survey, Enteromius toppini occupied the top of the food web. Consumer diets in Lake Nyamithi between the various surveys preferred different food sources. During the 2017 survey, C4 plants (32%) and macroinvertebrate (36%) were relatively equally preferred by consumers. Conversely, in the 2018 survey consumers preferred a larger variety of dietary constituents,

such as, detritus (30%), aquatic vegetation (29%) and macroinvertebrates (29%) while fish were the least consumed food source (12%). Strontium isotope ratios in C. gariepinus indicated movement between the various aquatic systems within the PRF. Moreover, the 87Sr/86Sr ratios between the Usuthu River and Lake Shokwe were very similar indicating the movement between the two systems while the large variation in 87Sr/86Sr ratios in the Phongolo River indicated C. gariepinus moved between the Phongolo and Usuthu River (due to slight overlap in signature) as well as the Phongolo River tributaries. Mercury in freshwater systems across South Africa has been thoroughly studied due to the many coal powerplant and gold mines in Southern Africa. The final biotic assessment was to determine Hg bioaccumulation and biomagnification in the biota from the various aquatic systems within the PRF. Furthermore, human health risks associated with the consumption of fish muscle were also determined. Sediments from the systems within the PRF had detectable Hg concentrations apart from Lake Nyamithi. These concentrations were all positively correlated with TOC. Aquatic biota from the various aquatic systems indicated detectable Hg concentrations while predatory fish bioaccumulated the highest Hg concentrations in muscle tissue. Similar to the positive relationship between Hg in sediment and TOC, Hg in fish had positive relationships with fish length and trophic position. With its positive relationship with trophic position, trophic magnification factors (TMF) were determined and indicated biomagnification of Hg through the food webs in all aquatic systems apart from Lake Nyamithi which indicated bio-dilution of Hg. Moreover, as fish from the PRF form an integral part of the diet of the local communities in and around the floodplain, the consumption of fish from these aquatic systems could be alarming. However, human health risks indicated low to no risks associated with the consumption of Hg-contaminated fish. The present study found that the environmental quality of the Usuthu River and its associated floodplain lake, Lake Shokwe were higher than the Phongolo River and Lake Nyamithi. This coincided with the natural flooding regime of the Usuthu River together with less of an impact from the upper catchment anthropogenic activities. This subsequently influenced the aquatic biota from these downstream aquatic systems while the high diversity of the HF survey indicated the importance of hydrological connectivity in riverine floodplains. The hydrological connectivity in the PRF further indicated the importance of the natural flooding of the Usuthu River and its role in the structure and functioning of the downstream aquatic systems. Therefore, with its strong flows and high hydrological connectivity between the rivers and associated floodplain lakes, enhances the possibility of fish movement within the lower PRF.

More importantly, Hg analyses indicated detectable Hg concentrations in aquatic biota as well as biomagnification, however, the study found no potential human health risk associated with the consumption of fish from the PRF. The 87Sr/86Sr ratios in the otoliths of C. gariepinus indicated the movement between the various aquatic systems within the PRF. This furthermore indicated the possibility of the Usuthu River serving as a refuge area for aquatic biodiversity of the lower PRF.