Evaluation of a diatom-based index for acid mine drainage impacted permanent pans in the Mpumalanga Highveld region

The Mpumalanga Highveld region features a variety of wetlands, including permanent depressional wetlands such as grass pans, reed pans, open-water pans, and salt pans. The region also hosts the Highveld coal field, South Africa's second largest producing coal field. Increased mining and agriculture present the most serious threats to wetland ecosystems in the Mpumalanga Highveld. These wetlands are often used to store acid mine drainage (AMD) generated by coal mining activities. AMD contamination affects surface and ground water resources and agricultural lands in the vicinity, disrupting the growth and reproduction of aquatic and terrestrial organisms. A water quality index based on diatoms was created in 2017 to determine the impacts of AMD on permanent pans in the region. Diatoms (Bacillariophyceae) are a ubiquitous group of algae with well-known ecological preferences and a history of successful use in biomonitoring of aquatic ecosystems. Being that natural variability among the pans may compromise the accuracy of diatom indices, indices sensitive to acid mine drainage (AMD) impacts were previously produced for a range of pan types. These indices were able to distinguish diatom communities present in non-impacted pans from communities in AMD impacted pans. However, indices accounting for natural variability perform better than those not accounting for natural variation and initially sampling was only done during the dry season (from May to August). The aim of this study is to evaluate the performance of diatom indices created for different AMD-impacted pans in South Africa's Mpumalanga Highveld region. In order to assess the performance of the previously developed diatom index, it was tested on a validation dataset of diatoms sampled from pans impacted by AMD-disturbed and non-AMD disturbed pans (i.e. pans ranging from least disturbed by coal mining to highly disturbed). Epiphytic diatoms were sampled from reed, salt, and grass pans during both the wet and dry seasons. The AMD-multimetric index scores obtained for the pans were compared to scores of riverine indices as diatom-based indices that are frequently used in South Africa as several riverine indices have been recommended for biomonitoring in wetlands by reason that the majority of indices include cosmopolitan taxa with well-established correspondence to environmental parameters. The present study found that factors that influenced the composition of diatom communities included the degree of AMD-disturbance, pan type and seasonal change. Similarity Percentage (SIMPER) analysis was used to determine the intra- and infra-group similarity among the samples resulting from three factors: disturbance level (AMD), pan type and season. Canonical Correspondence Analysis (CCA) was used to ascertain the relationship between diatom assemblages and measured environmental variables and Pearson's correlation coefficient was calculated to establish whether significant correlations exist between calculated index scores and water quality variables. For all indices tested, accuracy was determined as the percentage of correctly classified instances (CCI) Cohen's Kappa coefficient (Κ) was used as a measure of the agreement between indices in rating ecological condition. Diatom assemblages showed strong correspondence with water quality variables which signify AMD-pollution. Higher AMD-MMIs scores were observed in non-impacted pans and low scores in impacted sites, which indicates proper correspondence with heavier coal mining impacts. The indicated conditions of reference pans were poor according to riverine indices, which were less successful in discriminating between reference sites and impacted sites. Riverine index scores correlated positively with indicators of AMD-disturbance. AMD-MMIs displayed strong negative correlations with AMD indicators and were more accurate in determining AMD-disturbance level than riverine indices The index was explicitly developed for the indication of AMD and specifically for depressional wetlands and similar indices may be designed for the various other wetland types in the region.

Grantor dc:publisher

North-West University (South Africa)

Year dc:date.issued

2021

Author dc:creator

• Erasmus, Melanie

Advisor dc:contributor.advisor

• Taylor, J.C.

• Ecological assessment
• Diatom index
• Acid mine drainage
• depressional wetlands

Language dc:language.iso

en

Identifier URI

0000-0003-3402-5030

OAI identifier oai:identifier

oai:repository.nwu.ac.za:10394/37999