Turnover of benthic macroinvertebrates along the Mthatha River, Eastern Cape, South Africa: implications for water quality bio-monitoring using indicator species

ABSTRACT

The ecological integrity of the Mthatha River is increasingly under threat from climate change and anthropogenic impacts of contamination by acidification, organic pollutants, industrial and agricultural chemicals, and waste water. This study assessed the ecological state of the river and the impact of water quality on benthic macroinvertebrate communities from 2015 to 2016. Faunal specimens were sampled and environmental variables measured at four sites that vary in topography and human impact along the river. A total of 2980 specimens were recorded at sites, and classified into 12 orders, 46 families and 71 species. Species composition and distribution were directly or indirectly affected by season, physico-chemical variables and habitat heterogeneity. CCA results showed that electrical conductivity, pH, temperature, dissolved oxygen and vegetation accounted for species distribution patterns. Both habitat-restricted and common/widespread species can be used as indicators for assessing the quality of the Mthatha River system. Habitat management along the river should aim at preserving sufficient indigenous vegetation, especially during summer when biotope requirements are optimal. The practicality of using species-level surveys on a routine basis can be challenging. It might be advisable to select a few indicator species identified in this study and to monitor them as surrogates for the whole community.

KEYWORDS:

• Benthic macroinvertebrates
• abundance
• richness
• environmental variables

Introduction

The deterioration of surface water quality globally has been attributed to increasing impacts of climate change and anthropogenic pressures (Jun et al. 2016). Habitat loss and degradation, invasive species, overharvesting, chemical and organic pollution remain the main threats to freshwater biotic systems, resulting in species losses and/or extinctions (Niba and Mafereka 2015). South Africa is a dry country with low, irregular and seasonal rainfall patterns, resulting in a broad range of aquatic ecosystems adapting to different water quality regimes and flow characteristics (DWAF 1996).

Many South African rivers, including the Mthatha River have been deteriorating in quality in the past few decades, largely due to an increase in industrial, agricultural and domestic activities within the catchments. Such activities alter its physico-chemical characteristics and ecological integrity (Alvarez-Cabria et al. 2010; Simaika and Samways 2011; Konig et al. 2014). Since different benthic taxa tend to exhibit different tolerance levels to individual environmental parameters, water of suitable quality is essential to maintain healthy populations of benthic fauna (Dallas and Day 1993; Fatoki et al. 2001; Dallas 2004). Many studies have used family-level analyses of benthic macroinvertebrates to assess river health condition. In this study, we propose that a species-level assessment is needed in order to obtain a better interpretation of the ecological state of selected sections of the Mthatha River.

The objectives of this study were to identify the benthic macroinvertebrate assemblage along the Mthatha River and its response to measured site variables, and make recommendations for monitoring the quality of the river based on findings of the study.

Materials and methods

Study area

The study was conducted in shallow sections of the Mthatha River at the following sites: Headwaters (HW: 31° 30′32.19″ S, 28° 23′ 47.57″ E), Langeni Forest (LN: 31°29′ 35.14″S, 28° 28′ 36.65″E), Kambi Valley (KM: 31° 28′ 17.76″S, 28° 36′ 53.59″E) and Ntsaka (NT: 31° 41′ 08.49″ S, 28° 49′13.64″E) (Figure 1).

Figure 1. Map showing location of sampling sites at Headwaters (H), Langeni (L), Kambi (K) and Ntsaka (NT) along the Mthatha River, Eastern Cape, South Africa.

The geology of the catchment comprises mudstones and sandstones, with scattered deposits of alluvium in some valleys. Soils in the catchment are moderate to deep and vary between sandy loam in the upper section and clayey loam in the section lower down the river, with extensive pine plantations in the Headwaters (Fatoki et al. 2001).

Site selection and description

Using a 30 m transect, four a-priori-selected sites along the Mthatha River were mapped out. Each site measured 25 m in length by 2 m (1 m on land and 1 m into water) wide, separated apart by at least 5 m.

The Headwater (HW) site was minimally disturbed, and characterized by clear, fast flowing, nutrient-free water with medium-sized boulders which form riffles. Marginal forest and riparian vegetation formed a canopy that reduced light penetration. Langeni Forest (LN) site had a wide range of habitats types, with river flow characterized by runs. Riparian vegetation was dominated by sedges (Cyperus papyrus L.), herbs (Setaria sp.), grasses (Ludwigia octovalvis Jacq.) and forbs. This site was moderately disturbed by human activity such as grazing and logging from a nearby saw mill. Riparian vegetation at Kambi Valley (KM) site comprised of alien plants, e.g. black wattle Acacia mearnsii, Lantana sp., Eucalyptus sp. and indigenous herbs. This site was highly disturbed by water extraction, grazing and sand mining activities that altered water flow, resulting in high sediment loads at this section of the river. Habitat at Ntsaka (NT) site comprised a sandy substrate, with a few species of emergent and submerged vegetation (e.g. chara sp. and Commelina Africana, respectively). The site was accessible to intensive human activity such as washing, water collection and grazing by cattle from the surrounding Ntsaka local community.

Benthic macroinvertebrate sampling

According to Buchwald (1992) and Phiri et al. (2012), macroinvertebrates may reside on or within sediments, or be associated with aquatic vegetation. Semi-quantitative surveys were done in 2015–2016 by sampling benthic fauna at sampling units (SUs) with a net once a month throughout the year between 11 am and 1 pm on clear sunny days, using an adapted version of the South African Scoring System (SASS 5) (Dickens and Graham 2002).

Samples were taken within 5–10 min, using the kick-method in which rocks and other benthic material were disturbed by foot, and faunal specimens allowed to follow the water current downstream into a soft 1 mm mesh, 30 cm × 30 cm net. The contents of each sample were washed to the bottom of the net and tipped into a water water-filled tray. Benthic macroinvertebrate specimens were then separated from organic and mineral matter, counted (after preliminary identification to the lowest taxonomic level possible), and released back into water. Where specimens could not be identified in the field, subsamples were preserved in 70% ethanol, and transported to the laboratory for further identification using voucher specimens, a Zeiss dissecting microscope (Model STEMI DV4) and with the aid of field guides by Hart et al. (2003), Appleton (2002), De Moor et al. (2003) and Nelson (2007). Specimen identities were verified by a specialist at the Albany Natural Science Museum, Grahamstown, South Africa.

Measurement of environmental variables

Measurements of physical, chemical and biotic variables that were hypothesized to affect benthic macroinvertebrate composition and distribution were made thought out the one-year sampling period from 2015 to 2016 at each of the SUs. Measurements commenced within 10–15 min of faunal specimen sampling. Electrical conductivity (CND), acidity (pH), dissolved oxygen (TDO) and temperature (TEMP) were measured in situ using a portable multi-parameter meter (Model Hach HQd). Plants within the SUs were classified into broad vegetation types such as aquatic vegetation (Aqua Veg) (plants in the river channel, partly or fully submerged), riparian vegetation (Rip Veg) (grasses, reeds, shrubs and sedges on the water's edge) and algae (Algae) (isolated and also on stones), according to Gerber and Gabriel (2002). Vegetation was estimated as percentage of total surface area of SU covered by these plants.

Data analysis

To describe the variation in environmental variables, means and ranges of all measured variables were calculated for the study sites during the sampling period. The statistical software programs DIVERSE and CLUSTER in PRIMER v. 6 (Clarke and Warwick 2001) were used to determine indices of diversity and classification of species data, respectively. Benthic species × SU data matrices were fourth root-transformed to balance rarer and common species. The Bray–Cutis measure of similarity (Bray and Curtis 1957) was then applied to the data to generate SU similarity matrices that were fused successively using group average linking.

Results describing patterns obtained using clusters were represented by a dendrogram. SU clusters were confirmed using a non-metric multi-dimensional scaling (NMDS) ordination plot. Ordination techniques in correspondence analysis (CA) and canonical correspondence analysis (CCA) from the computer software package CANOCO (Ter Braak 1988; Ter Braak and Smilauer 2002) were also used to describe community trends in faunal data and to identify environmental variables that best explained species distribution patterns at study sites. CCA is a direct gradient analysis technique that uses multiple regressions to select linear combinations of environmental variables that account for most of the variation in faunal distributions.

Results

Overall taxon richness

A total of 2980 benthic macroinvertebrate individuals were recorded, and were classified into 12 orders, 46 families and 71 species (Table 1). The orders Lepidoptera and Platyhelminthes were represented by one family and one species each while the Ephemeroptera was the richest order with seven families and 12 species. The family Baetidae (Ephemeroptera) had the most species and individual counts. Castanophleblia sp. (Leptophlebibidae) had the highest population count and occurred mainly at the Headwater and Langeni SUs like other members of the order.

Table 1. Taxonomic profile of benthic macroinvertebratess recorded at Headwaters, Langeni, Kambi and Ntsaka in 2015–2016.

				Headwaters			Langeni			Kambi			Ntsaka
Order	Family	Species	Codes	HW1	HW2	HW3	LN1	LN2	LN3	KM1	KM2	KM3	NT1	NT2	NT3	Total
Odonata	Aeshnidae	Aeshna sp.	Aesh	8	5	7	15	29	26	0	3	2	0	2	0	97
		Anax sp.	Anax	1	0	1	3	5	1	0	0	0	0	0	0	11
		Anaciaeschra motivagans	Amon	0	1	0	0	0	0	0	0	0	0	0	0	1
	Gomphidae	Ceraptogomphus pictus	Cpic	6	2	6	14	12	32	24	15	8	1	0	0	120
		Ceraptogomphus sp.	Cera	0	0	1	0	0	0	0	0	0	4	0	0	5
		Notogomphus praetatius	Nprae	2	2	0	2	0	0	0	3	0	0	0	0	9
	Libelluidae	Brachythemis leucostica	Blue	0	0	0	0	0	3	3	1	0	12	4	4	27
		Brachythemis lacustris	Blac	0	0	0	0	0	0	0	0	3	6	5	0	14
		Trithemis dorsalis	Tdor	0	0	0	0	0	0	0	0	0	2	0	0	2
	Coenogrionidae	Pseudagrion sp.	Pseu	0	0	0	0	0	0	0	0	0	1	2	5	8
	Synlestidae	Chlorolestes sp.	Chlsp	0	0	2	0	0	0	0	0	0	0	0	0	2
	Chlorocyphidae	Chlorocypha sp.	Chlcy	0	0	0	0	0	0	0	0	0	0	1	0	1
	Synlestidae	Cholrolestes fasciatus	Cfas	0	0	1	0	0	0	0	0	0	0	0	0	1
Decapoda	Potamonautidae	Potamautidae sp.	Pota	4	15	15	2	1	6	0	2	0	0	0	0	45
Plecoptera	Perlidae	Neoperla sp.	Neop	17	13	10	23	47	45	8	34	47	1	2	2	249
	Notonemouridae	Afronemoura amatolica	Aama	0	0	1	0	0	0	0	0	0	0	0	0	1
		Alphaniceropsis tubularies	Atubu	10	4	1	0	0	0	0	0	0	0	0	0	15
Ephemeroptera	Tricorythidae	Tricorythus sp.	Trico	66	46	27	40	43	30	2	4	5	0	0	0	263
	Leptophlebidae	Castanophlebia sp.	Cast	26	24	45	39	31	30	4	15	8	54	43	19	338
		Euthraulus sp.	Euth	0	1	3	9	5	0	0	0	0	8	17	4	47
	Heptageniidae	Afronurus sp.	Afron	1	9	8	51	44	34	0	1	0	0	0	1	149
	Caenidae	Afroceania sp.	Afroc	7	10	15	8	16	10	6	9	2	0	0	0	83
	Prosopistomatidae	Prosopistoma sp.	Pros	0	0	0	0	0	0	0	1	2	0	0	0	3
	Oligoneuridae	Oligoneurus sp.	Olig	0	0	0	0	3	0	0	0	0	0	0	0	3
	Baetidae	Demoreptus sp.	Demo	15	12	17	2	3	10	0	1	2	1	0	0	63
		Pseudocloeon glaucum	Pglac	6	9	2	15	15	5	2	0	2	27	16	18	117
		Rheoptilum sp.	Rheo	12	5	6	3	2	7	2	1	1	1	2	1	43
		Acantiopsis erepens	Aerep	14	11	11	17	14	15	0	6	8	5	7	6	114
		Cloeon sp.	Cloe	16	3	4	2	2	5	4	9	14	16	13	17	105
Annelida	Oligochaeta	Lumbricus variegatus	Lvar	2	1	0	0	0	0	0	0	0	0	0	0	3
	Hirudinea	Oosthuizobdella stuhlmanni	Ostuh	0	0	0	0	0	0	0	0	0	18	34	23	75
		Asiaticobdella funeticta	Afun	1	0	1	0	0	0	0	0	0	0	0	0	2
Coleoptera	Elmidae	Pachyelmis convexa	Pcon	2	1	1	1	1	0	0	0	0	0	0	0	6
		Ctenelmis sp.	Cten	6	0	4	6	9	3	0	6	4	6	7	3	54
		Elminae sp.1	Elmi1	0	0	1	2	1	3	0	0	0	1	1	0	9
		Elminae sp. 2	Elmi2	0	0	0	0	3	0	0	0	0	0	0	0	3
	Helodidae	Helodidae sp.	Helo	1	0	1	0	0	1	0	0	0	0	0	0	3
	Dyticidae	Hydaticus sp.	Hyda	0	1	0	1	0	0	3	0	0	0	0	0	5
	Gyrinidae	Dineutus sp.	Dine	0	0	0	0	1	0	0	0	0	0	0	0	1
	Psephanidae	Afrobrianax ferdyi	Aferr	0	0	0	0	1	0	0	1	1	0	0	0	3
Diptera	Athericidae	Athericid sp.	Athe	0	2	4	3	3	0	0	2	0	0	0	0	14
	Tipulidae	Tipulia pompasa	Tpom	1	0	0	0	1	0	0	0	0	0	0	0	2
		Limnophila sp.	Limn	10	5	1	1	3	0	6	3	2	1	0	0	32
	Simullidae	Simullum letabum	Slet	28	5	2	20	6	11	3	26	12	22	25	13	173
		Simullum adersi	Sade	0	0	0	2	0	0	0	0	0	5	4	7	18
		Simullum vorax	Svor	0	0	0	1	0	0	0	0	0	8	3	6	18
	Chironomidae	Chironomini sp.	Chir	1	0	0	1	0	0	3	0	1	5	10	7	28
	Ceraptogonidae	Bezzia sp.	Bezz	1	4	1	1	0	0	0	0	0	0	0	0	7
	Culicidae	Malaya sp.	Mala	0	0	0	0	0	0	0	0	0	0	0	3	3
	Tabanidae	Tabanus sp.	Taba	2	0	3	0	0	0	0	0	0	0	0	0	5
	Mucidae	Muscid larvae	Musc	0	0	2	0	0	0	0	0	0	0	0	0	2
	Blephariceridae	Elporia sp.	Elpo	2	0	0	0	0	0	0	0	0	0	0	0	2
Tricoptera	Hydropsychidae	Cheumatopsyche afra	Cafra	5	0	1	3	5	0	0	2	0	5	1	0	22
		Cheumatopsyche thomasseti	Ctho	0	0	0	0	1	0	1	2	1	4	0	0	9
		Hydropsyche longifurca	Hlong	56	20	43	35	23	20	24	45	64	61	33	15	439
		Hydropsyche sp.1	Hydro1	0	0	0	0	0	0	0	1	2	0	1	0	4
	Leptoceridae	Leptocerus sp.1	Lept1	3	4	8	0	1	0	2	0	0	0	0	0	18
		Leptocerus sp.2	Lept2	0	2	0	0	0	0	0	0	0	0	0	0	2
	Philopotamidae	Dolophilodes urceolus	Durc	0	0	0	0	0	0	0	0	0	1	0	0	1
Hemiptera	Naucoridae	Laccocoris sp.	Lacc	0	0	0	0	0	0	0	3	1	0	0	2	6
	Belostomatidae	Appasus sp.	Appa	0	0	0	0	0	0	0	0	0	0	0	2	2
	Corixidae	Trichocorixia sp.	Trich	0	0	0	0	0	3	11	1	4	0	2	12	33
		Micronecta sp.	Micr	0	0	0	0	0	0	0	1	0	0	0	0	1
	Notonectidae	Enitthares sp.	Enit	0	0	0	0	1	0	0	0	0	0	0	0	1
		Notonecta sp.	Noto	0	0	0	0	0	0	0	0	0	0	0	4	4
	Vellidae	Rhagovelia sp.	Rhag	0	0	0	2	1	1	0	0	0	0	0	3	7
	Aphelocheiridae	Aphelocheirus sp.	Aphe	0	0	0	0	0	0	3	2	0	0	0	0	5
Platyhelminthes	Planaria	Dugestia sp.	Duge	0	0	1	0	0	0	1	0	0	4	3	2	11
Lepidoptera	Crambidae	Crambid larvae	Cram	0	1	0	0	0	0	0	0	0	0	0	0	1
Mollusca	Lymnaeidae	Lymnaea columella	Lcol	0	0	0	0	0	0	0	0	0	0	0	2	2
	Ancylidae	Burnupia sp.	Burn	0	0	0	0	0	0	3	0	1	3	0	1	8
Totan number of individuals				332	218	257	324	333	301	115	200	197	283	238	182	2980
Total number of species				30	28	35	30	32	22	19	28	24	27	24	26	325
Shannon diversity index (H')				2.8	2.7	2.8	2.7	2.7	2.6	2.5	2.6	2.3	2.6	2.5	2.6	31.4
Pielou evenness index (J')				0.8	0.8	0.7	0.8	0.8	0.8	0.8	0.7	0.7	0.7	0.8	0.8	9.2

Spatial distribution

Higher benthic macroinvertebrate individual counts were recorded at the Headwaters (HW) and Langeni (LN) than at Kambi (KM) and Ntsaka (NT). Langeni SUs had the highest population count (n_ind = 958) while the highest species numbers were recorded at the Headwaters (n_sp = 93). The lowest number of species (n_sp = 19) and individuals (n_ind = 115) across all SUs occurred at Kambi (SU KM1). Trends in species diversity (Shannon H’) and evenness (Pielou J’) across sites are presented in Table 1. Six species, Neoperla sp., Castanophlebia sp., Rheoptilium sp., Cleon sp., Simulium letabum and Hydropsyche longifurca, were recorded across all site SUs. Only one macroinvertebrate species (n_ind = 1–10) belonging to the habitat-tolerant family Chironomidae was recorded during the study.

Temporal distribution

Species richness and abundance varied across sites during the study period. Aeshna sp. and Neoperla sp. were recorded throughout the study at Langeni, while Potamonautidae sp., Tricorythus sp. and Afroceania sp. occurred every month at the Headwaters. Hydropsyche longifurca occurred each month at Ntsaka (Table 2). Highest monthly species and individual counts were recorded in April and September at Langeni and Ntsaka, respectively. Seasonal species counts showed that 21% (mean 15, range 10–18) of the total number of macroinvertebrate species recorded during the study occurred in spring and 17% (mean 12, range 6–17) occurred in summer.

Table 2. Temporal distribution of benthic macroinvertebrate taxa during the sampling period (2015–2016) across sites: H = Headwaters, L = Langeni, K = Kambi and N = Ntsaka Only species with population counts above two individuals per sampling unit have been included in the table.

Order	Family	Species	January	February	March	April	May	June	July	August	September	October	November	December
Odonata	Aeshnidae	Aeshna sp.	HL	HL	L	HL	HL	HL	L	L	L	L	L	HL
		Anax sp.		L	L	L	L				L
	Ghomphidae	C. pictus	HLK	HLK	LK		HK	HLK	LK	LK	LK	LK		HL
	Libellulidae	B. leucosticta		N	KN	KN	KN	N	N		N
		B. lacustris						K	K	N	N	KN
	Coenagrionidae	Pseudagrion sp.		N	N				N	N	N		N	N
Decapoda	Potamonautidae	Potamonautidae sp.	HL	HL	H	H	H	H	H	HL	H	HL	HL	HL
Plecoptera	Peridae	Neoperla sp.	HL	LK	LKN	HLK	HLK	HLK	HLKN	HLK	HLKN	HLKN	L	HLK
	Notonemouridae	A. tubularies						H	H	H	H
Ephemeroptera	Tricorythidae	Tricorythus sp.	HL	HL	HL	HL	HLK	HLK	HLK	HL	HL	HL	HK	HL
	Leptophlebidae	Castanophlebia sp.	HLN	HKN	HLN	LN	LN	HLN	HLKN	HLKN	HLKN	LK	N	HLKN
	Heptagenidae	Afronurus sp.	L	HL	L	HLK	LK	HLK	HLK	HLK	HLK	HLK	HL	HL
	Caenidae	Afroceania sp.	H	H	H	HK	HK	HLK	HLK	HLK	HLK	HK	HL	H
	Baetidae	Demoreptus sp.	HLK	HL		HL			HL	HLK	H	HL	HL	HLK
		Acantiopsis sp.	HKN	HLN	HN	L	HLKN	HLKN	HLKN	HLKN	HLKN	HKN	HLKN	HKN
		Cleon sp.	HLKN	HKN	KN	N			HLKN		HLN	HKN	HLKN	N
		Rheoptilium sp.	HL	HL	H	H		HLN				HLK	HLKN	HLKN
		P. glaucum	LN	HLN	HLN	HLKN		HL			HLK		HLKN
Coleoptera	Elmidae	Ctenemis sp.	H	N	HL		HLN	H	HL	HL	HLKN	K		HLK
		Elminae larvae				L				L	L	L
Diptera	Athericidae	Athericid sp.	H			HL	L			LK	LK	L		H
	Tipulidae	Limnophilia sp.		L	HK	HLK	HL	HK	HK	HL	HK	HK		K
	Simulidae	S. letabum		K	HK	HLKN	HLK	HLK	HLKN	HLK	HLKN	HN	HN	HN
	Chironomidae	Chironomini sp.	KN	N				KN	KN	KN			N	N
Annelida	Oligochaeta	O. stuhlmanni		N	N	N	N	N	N		N	N	N	N
Trichoptera	Hydropsychidae	H. Longifurca	HLN	HLKN	HKN	KN	HLKN	HLKN	HLKN	HLKN	HLKN	HLKN	HN	HLKN

Similarity of benthic macroinvertebrate distribution at sites

NMDS results showed that all SUs at Headwater and Langeni were over 65% similar, and showed closer similarity to Kambi SUs KM2 and KM3 at 58% similarity, while SUs at Ntsaka were 70% similar, and distinct from the other SUs (Figure 2(a,b)). Out of the 71 species recorded during the study, 36 species where distributed as follows: 7 occurred at all 12 SUs, 11 were restricted to Headwater SUs, while an additional 5 species (Pachyelmis convexa, Helodidae sp., Hydaticus sp., Tipulia pompasa and Bezzia sp.) were unique to Headwaters and Langeni SUs. Four (Oligoneurus sp., Elmidae sp. 2, Notonecta sp., Rhagovelia sp.), three (Prosopistoma sp., Micronecta sp., Aphelocheirus sp.) and five (T. dorsalis, Pseudagrion sp., Malaya sp., D. urceolu, Appasus sp., Chlorocypha sp.) were restricted to Langeni, Kambi and Ntsaka SUs, respectively.

Figure 2. Dendrogram showing (a) Bray–Curtis similarity and (b) NMDS obtained using group-average linking from matrices on fourth-root transformed benthic macroinvertebrate data for all sampling units at Head waters (HW), Langeni (LN), Kambi (KM) and Ntsaka (NTS) in 2015–2016.

Figure 3. CCA tri-plot for all sampling units depicting relationship between sampling units (circles), species (triangles) and measured environmental variables (arrows). Axis 1 is horizontal and axis 2 vertical. Species code names are given in Table 1. TEMP = temperature, CND = electrical conductivity, TDO = dissolved oxygen, Aqua veg = aquatic vegetation, Rip veg = riparian vegetation.

Environmental variables at sites

Measured environmental (physico-chemical and biotic) variables across site SUs during the study period are summarized in Table 3. Temperature was lowest in winter at Kambi (mean 14, range 9.5–19.5 °C) and highest in summer at the same site (mean 23 °C, range 17.4–28.9). PH varied from acidic (pH 6.2) at the Headwaters in summer to alkaline (pH 9.2) in spring at Kambi and winter at Langeni. Even though dissolved oxygen (TDO) showed little variation across sites, it was high in winter at Kambi. Mean variation of electrical conductivity across sites was high, and ranged from 39 µS/cm (range 27–51) in autumn at the Headwaters to 1643 µS/cm (range 1442–1844) at Ntsaka in autumn. While all vegetation categories (aquatic, riparian and algae) were recorded in Ntsaka, only riparian vegetation occurred in the other SUs across sites and sampling period.

Table 3. Environmental variables (mean, with range in parentheses) at SUs across sites: Headwaters (H), Langeni (L), Kambi (K) and Ntsaka (N).

	Acidity	TDO	TEMP	CND	Aqua veg	Rip veg	Algae
	(pH)	(mg/L)	(^οC)	(µS/cm)	% cover	% cover	% cover
H
Summer	7.4 (6.2–7.9)	8.6 (8.3–8.8)	21 (16.6–25.6)	68 (39.1–97.7)	na	2.5 (2–3)	na
Winter	6.9 (6.8–7.1)	8.2 (7–9.5)	14.6 (11.6–17.6)	153 (132–174)	na	10 (5–15)	na
Spring	7.7 (6.9–8.8)	8.1 (7.7–8.5)	20 (17.4–22.5)	90 (29–153)	na	2 (1–3)	na
Autumn	7.3 (7.01–7.6)	8.3 (8.1–8.54)	19 (18–20)	39 (27–51)	na	3 (1–5)	na
L
Summer	7.5 (6.9–8.2)	8.8 (8.6–9.22)	21 (15.5–26.8)	96 (60.1–133.3)	na	2 (1–3)	na
Winter	8.1 (7.1–9.2)	9.7 (9.1–10.3)	14.2 (10.1–18.2)	227 (220–235)	na	9 (3–15)	na
Spring	7.5 (7.8–8.7)	8.3 (7.8–8.73)	22 (19.4–24.6)	144 (41–247)	na	7.5 (5–10)	5 (1–9)
Autumn	7.6 (7.04–8.13)	8.5 (8.41–8.6)	19.3 (16.4–21.9)	84.4 (37.8–131)	na	3 (1–5)	na
K
Summer	7.7 (7.4–8.1)	8 (7.14–8.9)	23 (17.4–28.9)	142 (84.9–199.1)	na	3 (1–5)	2 (1–3)
Winter	7.1 (7.02–7.08)	10.2 (10–10.35)	14 (9.5–19.5)	370 (356–385)	na	10.5 (1–20)	na
Spring	8.9 (7.8–9.2)	8.2 (7.2–8.9)	22.8 (19.2–26.4)	281 (59.9–503)	na	3.5 (2–5)	2 (1–3)
Autumn	7.5 (7.1–7.9)	7.4 (7.1–7.8)	21 (18.7–23.2)	162 (48–276)	na	3 (1–5)	na
N
Summer	7.5 (7.02–7.9)	7.5 (6.99–7.92)	23 (19.4–28.4)	1320 (1118–1522)	6.0 (2–10)	8.5 (2–15)	9 (2–16)
Winter	7.5 (7.1–8.4)	8.1 (7.01–9.22)	15.3 (11–19.5)	1643 (1442–1844)	4 (3–5)	5 (2–8)	8 (5–11)
Spring	8.1 (7.01–9.19)	8.5 (6.85–10.2)	21.2 (19.2–23.2)	1020 (309–1731)	3.5 (2–5)	5 (1–9)	13 (6–20)
Autumn	7.62 (7.1–8.2)	7.1 (7.01–7.05)	22.9 (23–22.7)	1275 (1250–1300)	6.5 (1–10)	3 (1–5)	6.5 (1–10)

Response of benthic macroinvertebrates to measured site variables

Eigenvalues for CCA ordination tri-plot and species-environmental variable correlations were strong (Table 4). Although Monte Carlo permutation tests for axis 1 was not significant (F = 2.7; P > 0.05), tests for all four canonical axes (Global test) were significant (F = 2.5; P < 0.05). Cumulative percentage variance of species-environmental variable relations for axis 1 and 2 was 63.7%, with species clustered to the center of the CCA ordination tri-plot (Figure 3). This suggests that no single environmental variable fully explained species composition and distribution pattern across sites. However, intra-set correlation showed that pH, conductivity, TDO, temperature and some vegetation categories accounted for species composition and distribution across sites (Table 5).

Table 4. Weights of the first canonical correspondence analysis (CCA) for benthic macroinvertebrate assemblages recorded at Headwaters (H), Langeni (L), Kambi (K) and Ntsaka (NT) in 2015–2016. Variances explained by both axes are given. The results of the Monte Carlo permutation tests are shown for axis 1 and all four axes of the CCA ordination.

Statistic	Axis 1	Axis 2	All four axes
Eigenvalue	0.373	0.261	0.840
CPVSp^a	36.2	51.9
CPVS-Env^b	44.4	63.7
Sp-Env C^c	0.994	0.985
F-ratio	2.27 ^ns		2.52^s

Note: s = significant at p < 0.05; ns = non-significant.

^a Cumulative percentage variance for species data.

^b Cumulative percentage variance of species-environmental variable relations.

^c Species-environmental variable correlations.

Table 5. Correlation between measured environmental variables and benthic macroinverebrates at site sampling units in 2015–2016. CND = conductivity, TDO = Total dissolved oxygen, Aqua veg = aquatic vegetation, Rip veg = riparian vegetation.

	Intra-set correlation		Inter-set correlation
Variable	CCA axis 1	CCA axis 2	CCA axis 1	CCA axis 2
pH	0.7565	0.4745	0.7055	0.4634
CND	0.9765	−0.1222	−0.7067	−0.3625
TDO	−0.9089	−0.0138	0.09205	0.3325
TEMP	0.8672	0.4622	0.5352	0.3160
Aqua veg	0.9405	0.1835	−0.3615	−0.4866
Rip veg	−0.3481	−0.0418	0.6410	0.1294
Algae	0.9625	0.1885	0.9739	0.3843

Discussion

According to Shannon diversity (H’) and Pielou evenness (J’) indices trends, higher richness and abundance of benthic macroinvertebrates were recorded at the Headwaters, with 11 species (15% of the total number of species recorded) restricted to this site. Five additional species occurred only at Headwaters and Langeni sites probably due to optimal habitat conditions, resource availability and comparatively lower levels of anthropogenic impacts at these two sites (Wolmarans et al. 2014; Niba and Mafereka 2015). This is not surprising as relatively higher numbers of pollution-sensitive taxa, e.g. Ephemeroptera, Plecoptera, Trichoptera and Odonata (EPTO) occurred mainly at these sites. However, not all species or even families belonging to the EPTO group are highly sensitive (Dickens and Graham 2002). Baetis harrrisonii (Baetidae) is probably the most common mayfly in South Africa, occurring at virtually all benthic habitats. Very low counts of the Chironomini sp. (n_ind = 1–10) representing the habitat-tolerant family Chironomidae occurred at Langeni and Ntsaka SUs (down the Mthatha River) probably as a result of lower biotope quality at these two sites compared to other site SUs up the river.

Temporal trends in species richness and abundance

The use of benthic macroinvertebrates to evaluate water quality allows for detection of temporal episodes that occur during seasons and between sampling stations, which physico-chemical assessments alone cannot establish (Dickens and Graham 2002; Niba and Samways 2004). Thus, a combination of biological and physico-chemical methods as proposed by Duran (2006) is best suited for determining water quality at different seasons.

Overall lower macroinvertebrate richness in summer may be attributed to higher rainfall, resulting in increased river flow velocity especially in Kambi and Ntsaka (Table 3). The high electrical conductivity values recorded at these two sites probably resulted from increased run-off from agricultural, milling and other anthropogenic activities and could explain the overall low species richness and abundance trends here. On the other hand, higher species counts in April at Langeni may have been due to higher habitat complexity, shelter from predators and resource availability, while lower species counts at sites may have been associated with less optimal habitat conditions (Simaika and Samways 2011). Seasonal trends in macroinvertebrate occurrence along the Mthatha River will need to be studied further with a closer analysis of species life cycles (Niba and Mafereka 2015).

Effects of measured variables on species distribution

Benthic macroinvertebrate species populations respond differently to disturbance regimes and many species have different habitat requirements at each stage of their developmental life cycle (Simaika and Samways 2011). Most species in this study clumped at the center of the CCA ordination tri-plot suggesting that no specific environmental variable defined species distribution patterns observed. However, degrees of importance of measured variable gradients denoted by length of arrow on the ordination plot and their related intra-set correlation coefficients could be selected. These were electrical conductivity, pH, temperature, dissolved oxygen (TDO) and vegetation characteristics.

Electrical conductivity across sites ranged from 39 to 1643 µS/cm, and was within the range required for other freshwater bodies in South Africa (De Kock and Wolmarans 1998). This environmental variable gradient increased in value from 27 µS/cm at the Headwaters in autumn to 1844 µS/cm downstream in Ntsaka during winter (Table 3), and was consistent with the disturbance trends at sites that increased from Headwater to Ntsaka SUs. This variable occurred on axis 1 of the ordination tri-plot, and probably accounted for the occurrence and distribution of Pseudagrion sp, and Ceratogomphus sp at Ntsaka SUs.

Simaika and Samways (2011) have noted the importance of pH in structuring benthic macroinvertebrate assemblages in the naturally acidic Tsitsikama River, compared with the neutral to slightly acidic Mthatha River. Most freshwaters in South Africa are relatively well buffered and more or less neutral, with pH ranges of 6–8 (Thiere and Schultz (2004). Although minimum pH values were never lower than the minimum values of the pH range for African freshwater bodies (Holmes 1996), maximum values in this study (mean 8.9, range 7.8–9.2) occurred in spring at Kambi. The pH and temperature gradients in this study were positively correlated with axis 1 of the CCA ordination, and probably determined the composition and distribution of Trichocorixa sp., Hydropsyche sp., Brachythemis leucosticta, B. lacustris and Trithemis dorsalis at Kambi SUs.

Annual rainfall in South Africa ranges from 100to 1200 mm, and this variability is usually magnified in the hydrological cycle of streams and rivers (Eady et al. 2013). This phenomenon in turn plays a critically important role in determining the energy dynamics and therefore thermal requirements of benthic fauna (May 1976, 1984). Higher temperature ranges of 9.5 °C (winter) to 28 °C (summer) at Kambi and similar trends across sites (Table 3) probably influence biotope requirements and hence the assemblage structure of benthic macroinvertebrates during the study period.

Dissolved oxygen (TDO) is not only influenced by river flow velocity but also the mechanisms of storage, distribution and transport of fine sediments on the river bed (Miserendino and Masi 2010). TDO was generally higher at the Headwaters and Langeni SUs that were associated with riffles and run biotopes, respectively. This probably enhanced the recruitment of diverse benthic faunal communities, as a result of optimal habitat conditions (Kefford et al. 2010). Lower values of dissolved oxygen and therefore faunal richness were associated with Kambi and Ntsaka. It may be possible that as the Mthatha River flows downstream, more contaminants from the catchments get into the river.

Effects of vegetation

Vegetation plays an important role in aquatic ecosystems by providing habitat complexity, shelter, breeding sites, substrata for periphyton and abundant resources for many aquatic fauna (Ali et al. 2007; Valinoti et al. 2011). Thus, the replacement of indigenous plant species by exotics in riparian zones can affect the quality of available detritus and colonization by detritivores (Konig et al. 2014). Riparian vegetation may also play a role in protecting benthic fauna from the direct effects of UV radiation (Quinn et al. 2004). Furthermore, macrophytes produce oxygen through photosynthesis and so create more optimal biotope conditions (Tillin et al. 2008).

In this study, even though riparian vegetation was important across sites (especially at Headwaters and Langeni) and accounted for faunal distributions, aquatic and algae vegetation characteristics were positively correlated with axis 1 of the CCA ordination tri-plot in determining benthic macroinvertebrate assemblage composition and distribution at Ntsaka SUs.

Macroinvertebrate sensitivity

Both rare/habitat restricted and common/widespread benthic macroinverebrates were sampled during the study. Of the 71 species recorded across all SUs, seven species occurred across all SUs with population ranges of (n_ind = 1–338). Eleven species of which 4 (n_ind = 1) and 7 (n_ind = 1–10) were restricted to the Headwater SUs while four, three and five species were restricted to Langeni, Kambi and Ntsaka SUs, respectively (Table 1). These species can be used as potential indicators for assessing the quality and conservation value of such habitats along the Mthatha River.

Bio-monitoring implications

Habitat patch level management along the Mthatha River system should aim at preserving sufficient indigenous vegetation categories (riparian, aquatic and algae), especially during summer season when habitat requirements are optimal for most growth stages of benthic macroinvertebrate taxa. Higher family and species richness and abundance for pollution-sensitive members of the EPTO group were mostly recorded at Headwaters and Langeni SUs. The habitat at Kambi and Ntsaka, located lower down the river, need to be improved to approach natural conditions.

Both rare/habitat restricted and common/widespread benthic macroinvertebates identified in this study can be used as indicators for assessing the ecological integrity of the Mthatha River system. The practicality of using species-level surveys on a routine basis can be challenging, given the lack of available taxonomic expertise and time constraints. It might be advisable to select a few indicator species identified in this study and to monitor them as surrogates for the whole community.

Acknowledgments

The authors express their appreciation to Helen James, Freshwater Invertebrate Collection, Albany Natural Science Museum, Grahamstown, South Africa for assistance in species identification and for granting access to this facility. The running expenses of this study were covered by the National Research Foundation (NRF) through the Directorate for Research Development, Walter Sisulu University.

Disclosure statement

There is no potential conflict of interest by us.

Additional information

Funding

The National Research Foundation (NRF) through the Directorate for Research Development, Walter Sisulu University [grant number TTK13071221189/UID: 88085].

Notes on contributors

Augustine Niba

Augustine Niba is a senior lecturer and researcher at the Department of Biological & Environmental Sciences, Faculty of Natural Sciences, Walter Sisulu University, Eastern Cape Province, South Africa.

Selunathi Sakwe

Selunathi Sakwe is a masters student at Department of Biological & Environmental Sciences, Walter Sisulu University and researcher with the Ministry of Water Affairs and Sanitation, Nelspruit, Mpumalanga Province, South Africa.