Hydrobiologia 435: 167–175, 2000. © 2000 Kluwer Academic Publishers. Printed in the Netherlands.
167
Benthic macroinvertebrate communities of intermittent streams in the middle reaches of the Guadiana Basin (Portugal) A. M. Pires1 , I. G. Cowx2 & M. M. Coelho1,∗ 1 Centro
de Biologia Ambiental-Departamento de Zoologia e Antropologia, Faculdade de Ciências de Lisboa, Campo Grande Bloco C2-3◦ Piso - 749-016 Lisboa, Portugal Tel: + 351 21 7573141 (Ext: 1537). Fax: + 351 21 7500028. E-mail:
[email protected] 2 Hull International Fisheries Institute, University of Hull, Hull HU6 7RX, U.K. (∗ Author for corespondence) Received 25 February 1999; in revised form 22 May 2000; accepted 20 June 2000
Key words: macroinvertebrates, intermittent streams, environmental variables, canonical correspondence analysis (CCA)
Abstract The Guadiana River has an irregular hydrological regime, with severe droughts and floods, but little is known about how aquatic fauna respond to these natural events. Macroinvertebrate data and environmental information were collected at seven sites from three tributaries in the middle reaches of the Guadiana River, approximately every 3 months from April 1995 to April 1997. Despite considerable annual variation in discharge (related to duration of flood and drought periods), the number of macroinvertebrates found was consistently high. Diptera represented the major proportion of the benthic fauna (73.2%) followed by Ephemeroptera (10.3%), Coleoptera (4.1%) and Trichoptera (3.1%). Canonical correspondence analysis (CCA) was used to evaluate the relationships between taxa density and habitat variables. Generally, Plecoptera and Ephemeroptera were found in the upstream sampling sites. Wider and deeper sites were associated with the presence of Diptera and were least diverse. High values for both the Shannon-Wiener diversity index and the average score per taxon were usually found at upstream sites where Ephemeroptera and Plecoptera predominated. The data suggest that macroinvertebrates have a great capacity to recover rapidly from severe drought periods, both in terms of taxonomic diversity and number of individuals.
Introduction Historically, invertebrates have received considerable attention in the study of running water ecosystems (Cummins, 1992). In particular, relationships between macroinvertebrate community structure and environmental variables have been the subject of numerous investigations (e.g. Depiereux et al., 1983; Townsend et al., 1983; Wright et al., 1984; Ormerod, 1987; Ormerod & Eduards, 1987; Boulton & Lake, 1990, 1992a, b; Gower et al., 1994; Tate & Heiny, 1995). Invertebrate communities are also good indicators of water quality conditions (Resh, 1995; De Shon, 1995). Furthermore, studies of temporal variation in the community structure of streams (Fisher et al., 1982; Scrimgeour & Winterbourn, 1989) have
indicated that physical disturbances can be important determinants of community structure in lotic systems (Sousa, 1984; Resh et al., 1988). Intermittent streams, such as those found in the Guadiana Basin, Portugal, present a particularly severe physical environment. Streamflow variability has been identified as a major factor affecting other abiotic and biotic factors that regulate lotic macrozoobenthic processes and patterns (Poff & Ward, 1989). Boulton & Lake (1992b) considered that drying up is a predictable event in intermittent streams, although timing and duration vary on a fine scale, supporting the case that changes in community composition when flow resumes can be regarded as seasonal periodicity. The amplitudes in physical and chemical conditions, especially in receding pools, far
168 exceed those in permanent streams, and may be important determinants of the structure and composition of macroinvertebrate assemblages (Boulton & Lake, 1990). The aim of this study was to examine the benthic macroinvertebrate fauna of intermittent streams in the middle reaches of the Guadiana Basin and relate changes in the community structure to fluctuations in environmental conditions, particularly with respect to prevalent drought conditions. This forms part of a wider study to investigate the ecology of aquatic fauna in intermittent Mediterranean streams in southern Portugal (Pires et al., 1999).
were taken by kick sampling using a hand net (mouth: 30 × 20 cm; mesh 0.25 mm). Sampling was performed during 1 min, and care was taken to include all available microhabitats within a representative 50-m section of the stream. Macroinvertebrates were identified to family level, counted and expressed as the numerical percentage of each category in each period for each sampling site. A succession of dry years prior to 1995 reduced the river discharge, consequently in some periods no sampling was possible because some sites were completely dry. The beginning of 1996 was characterised by severe floods and it was not possible to sample until March. Environmental variables
Study area The Guadiana River is a typical Mediterranean river. It flows 810 km from its headwaters in Spain to its mouth in Portugal, with 150 km of the main river and many longer tributaries exclusively in Portugal, and 110 km of main river forming the border between the countries. It has an irregular hydrological regime, with severe droughts and floods, and is under increasing pressure for exploitation of water resources. The main rainfall period occurs between November and March, decreasing throughout the summer resulting in some tributaries of the river drying up or becoming a series of isolated pools. Seven sites (see Figure 1) on three tributaries in the middle reaches of the Guadiana, i.e. Xévora (A, B, C), Caia (D, E) and Degebe (F, g), were selected for this study. The rivers are different and coverage of the three is considered to be representative of the different river types in the region. Although all sites have some human pressure, the Xévora and the Caia are relatively unimpacted compared to the Degebe, which has higher organic pollution (see Pires et al., 1999, for details). Materials and methods Benthic macroinvertebrate sampling The benthic macroinvertebrate community was sampled approximately every 3 months from April 1995 to April 1997. This was considered adequate to detect recovery in the invertebrate communities following severe climatic events such as floods and drought. On each sampling occasion (April 95, July 95, October 95, March 96, April 96, July 96, October 96, February 97 and April 97), 3–5 samples of bottom fauna
Environmental data, describing river morphology and water quality, were also taken on each sampling occasion at each site. Repeated measurements of a number of environmental variables at several sites in different habitats over at least 2 years are essential to describe seasonal variation in the physico-chemistry of intermittent streams (Boulton & Lake, 1990). The data collected were: depth (cm), current velocity (m s−1 ), temperature (◦ C), conductivity (µS), oxygen (mg l−1 ), width (cm) and dominant substrate [(1) organic cover; (2) silty sand (1–2 mm); (3) sand (2–5 mm); (4) gravel (5–25 mm); (5) pebble (25–50 mm); (6) rock (50–100 mm); (7) cobble (100–250 mm); (8) boulder (>250 mm)]. The tendency of the site to dry up (i.e. absence of standing water over the river bed during the summer; 0 for sites which dry up and 1 for sites which maintained a flow) was considered as a characteristic of each site. Indices Several indices and scores were used to examine changes in the invertebrate species diversity with time and in relation to fluctuations in climatic events. These include the Shannon–Weiner Index (H 0 ) and the Biological Monitoring Working Party Score (BMWP). Diversity indices are mathematical expressions that use three components of community structure, namely richness (number of categories present), evenness (uniformity in the distribution of individuals among categories) and abundance (total number of organisms present), to describe the response of a community to the quality of the environment (Zar, 1984; Metcalfe-Smith, 1996). The most widely used diversity index is the Shannon–Wiener index (H 0 ) because it is stable in any spatial distribution and insens-
169
Figure 1. Sampling sites on the three tributaries in the middle reaches of the Guadiana basin: X´evora river (A, B, C); Caia river (D, E) and Degebe river (F, G).
itive to rare species (Ludwig & Reynolds, 1988). The higher the value of H 0 , the greater the diversity and, supposedly, the healthier the environment (MetcalfeSmith, 1996). However, according to Bendati et al. (1998), a few individuals evenly distributed among several species could give a relatively high index of diversity even though a habitat is grossly polluted. Despite this problem, the utilization of a diversity index can aid in data reduction, and, in combination with the other indicators of community structure (rich-
ness, evenness and abundance), allows interpretation of environmental condition and the visible effects on macroinvertebrate organisms. Kendall’s Coefficient of Concordance was used to test for differences between the diversity index (Zar, 1984). When significant variation was found, between sites or between periods, an approximate t-test was used to evaluate the cause for the differences (Zar, 1984). The Biological Monitoring Working Party (BMWP) method is widely used because organisms are identi-
170 fied to family level for uniformity, and families with similar pollution tolerances are grouped together. The ‘average score per taxon’ (ASPT), which simply refers to the total score divided by the number of scoring taxa, has frequently been applied to the BMWP score because it is independent of the number of taxa counted (Metcalfe-Smith, 1996). Data analysis Canonical correspondence analysis (CCA) was used to evaluate the relationships between taxa density and the environmental variables, and was performed using the program CANOCO (ter Braak, 1987–1992, 1990). CCA compares variation in community composition by constraining ordination axes to be linear combinations of environmental variables. The resulting taxa-environmental biplot is an ordination diagram in which taxa and sites are represented by points, and the environmental variables are represented by arrows. The arrows roughly orientate in the direction of maximum variation in value of the corresponding variable (ter Braak, 1986, 1987–1992; ter Braak & Verdonschot, 1995). A Monte Carlo permutation test with 9999 permutations was used to test whether species abundance was related to environmental variables on both the first axis Eigenvalue and the trace, the sum of all the Eigenvalues (ter Braak, 1987-1992; ter Braak & Verdonschot, 1995). The incentive for using a high number of permutations is to increase the power of the test (ter Braak & Wiertz, 1994). CCA is a powerful exploratory tool for simplifying complex data sets and has the advantage over methods of indirect gradient analysis of allowing integrated analysis of both species and environmental data. CCA also provides an effective method for analysing patterns of multiple environmental variables on invertebrate assemblages.
Results Despite the severe environmental conditions experienced in the intermittent streams of the Guadiana, the density of macroinvertebrates was often high (Figure 2). However, a major proportion of the benthic fauna (73.2%) comprised Diptera larvae, Ephemeroptera nymphs (10.3%), Coleoptera larvae (4.1%) and Trichoptera larvae (3.1%). Fewest invertebrates occurred in the Degebe River (Figure 2). Despite the reduction in the number of individuals for some sampling periods, the diversity index did not alter
significantly (P>0.05) (Figure 3). Significant differences (P0.05) were found between periods for each site. Although the Xévora River (sites A, B and C) had the highest abundance and diversity, for some periods site C had low values for the diversity index (Figure 3). Generally, a high proportion of Diptera, particularly Chironomidae and Simuliidae, corresponded to the lowest values for the diversity index (Figures 2 and3). Differences in the diversity index were particularly marked for site D on the Caia River, which was completely dry in July 1995 and October 1995, but the macroinvertebrate community recovered rapidly after the drought period. The flood during the winter of 1995/1996 (January– March) generally reduced the number and diversity of macroinvertebrates at all sites, particularly at the most upstream site (site A in the Xévora River). Generally, the BMWP index (Figure 3) decreased in a downstream direction. The highest values were at sites A (BMWP = 45), B (BMWP = 39) and D (BMWP = 39). The value for site A corresponded to October 1995, and appears to reflect the good water quality and conditions for this upstream site, even during the summer. The BMWP scores for site C (lower reaches of the Xévora River) were consistently lower (range 2–14) than the other sites on the river, but were similar to the Degebe River sites, and probably indicative of poor water quality found in the pool habitats during dry weather conditions. Dry periods coincided with reduced dissolved oxygen levels (as low as 2 mg l−1 compared to 8–12 mg l−1 in wet weather conditions) in water. The poor water quality conditions were reflected by the dominance of Diptera larvae in the aquatic invertebrate fauna in the lower reaches of the rivers, which in turn resulted in the low BMWP scores. ASPT scores similarly mirrored the water quality conditions indicated by the BMWP index (Figure 3), and were correlated (P
