A River Continuum From Stream To River Biology Essay

Published: Last Edited:

This essay has been submitted by a student. This is not an example of the work written by our professional essay writers.


Abiotic and biotic parameters were measured at eight sites along the river Tanat, a tributary of the river Severn, to identify the abiotic river continuum and to determine the effects of physical and chemical parameters on the distribution of macro-invertebrates. Results showed that channel morphology (width & depth) was significantly correlated to distance from source (km) as were a number of chemical parameters including temperature, conductivity and total hardness. Nutrient concentrations (NO3 & PO4) indicated no significant relationship with distance downstream and analysis of macro-assemblages based on Functional Feeding Group (FFG) yielded little difference between sampling sites. A Canonical Correspondence Analysis (CCA) based on the presence absence of taxonomic families and transformed environmental data found to be significantly correlated to taxa richness (r>0. 65, p<0.01) indicated that molluscs and crustaceans were predominately associated with downstream sites (sites 5-7) where total hardness and conductivity was highest. Stoneflies (Plecoptera) however were predominately associated with higher than average slope (m/km) and therefore located at lower order sites (sites 1-3).


Demand for the development of methodologies that allow accurate classification of stream environments has greatly increased in the last thirty years due to the enhanced role of river management and restoration (Raven et al., 2000). Increased global interest in examining rivers at a watershed level to assess the effects of global phenomenon such as climate change on freshwater resources in relation to water security and biodiversity has also played a major role (Frissell et al., 1986; Poff et al., 2003; VÓrÓsmarty et al., 2010). As a result a number of models have been developed to provide greater understanding of biotic and abiotic processes operating on different scales within four dimensional lotic environments (Vannote et al., 1980; Ward & Standford, 1983; Stanford & Ward, 1993; Junk et al., 1989).

The river continuum concept (RCC) for example, attempts to associate biological community function to longitudinal environmental gradients (Vannote et al., 1980). Vannote et al., grouped lotic communities spatially, by stream order and functionally, by feeding mode, in order to demonstrate how the structure and function of invertebrate community changes in response to downstream variations in environmental parameters. The Hyporheic Corridor Concept (HCC) however specifically examines the heterogeneity and functionality of large alluvial river floodplains which can provide habitat for groundwater and river boundary communities within the hyporheic zone due to upwelling of unconfined aquifers in the lower catchment (Stanford & Ward, 1993). The Serial Discontinuity Concept (SDC) examines the effects of empoundment on flow regimes and transfer of solutes along the river continuum (Ward & Stanford, 1983) however this was later extended to encompass the lower portion of the river system and the associated floodplain environments that rely on upstream resources for their maintenance and survival (Ward & Stanford, 1995).

These models however all have inherent limitations primarily associated with their assumptions. Statzner & Higler (1985) noted that as stream order provides not quantifiable measurement of the environment, it should not be used as indication of physical as it is done in the RCC, but rather an indication of a stream position within the overall catchment. Secondly, the confluence of tributaries and the presence of naturally occurring lakes along the main river channel may perturb the expected downstream habitat succession (Poole, 2002). The RCC is also concerned greatly with the equilibrium between energy flow and energy consumption however, Stazner & Higler (1985) noted that temporal variations in run off, discharge, sediment transport and auto/allochlonous inputs of organic matter are likely to produce seasonal or localised effects and concluded that physical parameters do not form a continuous gradient with distance downstream (Poole, 2002).

The main limitation of current models and strategies for habitat classification is that they often do not apply equal weighting to morphological, biological and chemical processes structuring aquatic ecosystems. The importance of sediment and fluvial dynamics in controlling the evolution channel geomorphology for example, in the past has been underestimated resulting in the need for large scale restoration works along stretches of the lower Mississippi River (Harmer et al., 2005). Additionally, models often operate on spatially and temporally isolated scales reducing their application to multi-dimensional processes. The RCC and the Serial Discontinuity Concept (SDC) for example, operate at stream to segment scales (103-102m) in the longitudinal direction whereas the Flood Pulse Concept and the HCC examines finer segment to reach scales (102-101m) in the lateral direction (Frissell et al., 1986; Junk et al., 1989; Stanford & Ward, 1993). Habitat classification strategies have therefore become increasingly focussed on multidimensional analysis resulting in the development of models centred on the River Discontinuum Concept (RDC) which examines fluvial environments in terms of hierarchical patch dynamics (Wu & Loucks, 1995). These models are capable of simulating three dimensional patch arrangement as well as the effects of changing longitudinal profiles on chemical constituents and biotic community structure (Poole, 2002).

This report aims to evaluate changes in abiotic and biotic parameters within the longitudinal plane of the River Tanat with reference to the RCC and the RDC in order to assess the assumptions made by both theories and to determine their applicability to classification of the river Tanat. Ability to classify river sections based on physical parameters will be considered in reference to the River Habitat Survey (RHS) and water chemistry will also be analysed in conjunction with physical parameters to investigate whether more comprehensive surveying strategies improve classification of river habitats. Finally, invertebrate presence absence will be examined with regards to environmental parameters in order to determine the dominant physical and chemical variables controlling the distribution of invertebrate taxa downstream.


Survey Area

Surveying was carried out along the river Tanat which is located in Northern Powys, Wales within the upper river Severn basin (Fig. 1). Originating from the Berwyn Mountains, the river Tanat forms one of the major tributaries of the river Vyrnwy which it joins north-east of Llansantffraid-ym-Mechain before flowing southeast across the border into north-west England. The two rivers have a collective catchment of 788km2 and predominately drain Ordovician and Silurian grit, silt and mudstones (Taylor & Brewer, 2001).

Figure 1: The North-west River Severn catchment showing topography and major towns within the region. The yellow outline highlights the river Tanat and the approximate area surveyed (Modified from the Geological Association, 2007)

The portion of the river Tanat surveyed encompassed the upper reaches, 0.2km from the rivers source to below the confluence of the river Tanat and the River Vyrnwy which was located between sites 6 & 7. Survey sites 1 & 2b were located along a first order stream whilst site 2a represents a second order stream which converged with the main channel a few metres downstream of the survey site (Table 1). These sites were located within the steep upper valley and in close proximity to one another. The river profile (Fig. 2) indicated an initial increase in the steepness of the landscape towards site 3, and then a plateauing from site 3 as the survey progressed downstream towards the middle and lower reaches.

2.2 Sampling

Physiological, chemical and biological data collection was conducted on 3-4th October 2009. At each site, depth (m) and flow velocity were measured at systematic intervals across the channel width (m) to determine discharge (m3 s-1). Temperature (oC), dissolved oxygen (%) and conductivity were measured as point samples whilst nutrient concentrations (NO3 & PO4) and pH were determined from in situ water samples although samples were also taken back to the laboratory for repeat analysis. Using these samples, total and calcium

Table 1: National grid references of survey sites and ordinance map derived site information

Survey Sites








Location (NGR)








Stream Order








Altitude (m)








DFS (km)








Slope (m/km)








hardness was also ascertained. Bed type was determined using the Wolman's Pebble Count Procedure (Wolman, 1954). An equipment list can be found in appendix 1. The presence absence of riparian vegetation and aquatic macrophytes as well as the identification of fish species at downstream sites was recorded however will not be examined within this report.

The present absence of macro-invertebrate taxa was determined using kick sampling which was conducted within the margins and central region of the channel at each of the eight sites and taken back to the laboratory for taxa identification.

Data Analysis

A multiple linear regression (p<0.05) was carried in Sigmaplot to determine whether changes in channel morphology and flow dynamics were correlated with distance from source (DFS). Bed type was omitted from statistics analysis due to the lack of qualitative data. To attempt classification of river sections based channel morphology and water chemistry, a combined Principal Component Analysis (PCA) was conducted on physical parameters that were significantly correlated to DFS and all chemical parameters excluding pH.

Invertebrate richness and the percentage of total richness represented by insects were determined based on the presence absence of taxa identified at each site. Average Scores Per Taxa (ASPT) was also calculated based on The Biological Monitoring Working Party (BMWP). Using existing literature (Mihuc, 1997 & Giller & Malmqvist, 1998), taxa were then divided into the four feeding guilds classified by Vannote et al., (1980) and the percentage of the total taxa represent by each feeding guild determined. These were examined on a site basis to identify any changes in dominant feeding guild with distance downstream.

The relationship between invertebrate community assemblage and abiotic factors was determined using a multivariate analysis technique adopted by Miserendino & Pizzolon (2003). Taxa richness was correlated to transformed (√x) physical and chemical parameters, using a Spearmans Rank Correlation Test (P<0.05) in Simaplot to determine which parameters varied significantly with the total number of taxa present. Parameters that exhibited no significant correlation to taxa richness such as depth, velocity, pH, dissolved oxygen and nitrate (NO3) concentrations were omitted as well as co-varying variables (r>0.65; p<0.01) such as altitude, distance from source, channel width, discharge, temperature and calcium hardness. A Canonical Correspondence Analysis (CCA) was then carried out using XLSTAT to determine the dissimilarity between macro-invertebrate communities identified at each of the eight sites and the effects of slope, conductivity, total hardness and phosphate (PO4) concentrations on the presence absence of taxa at a family level. Rare species and families exhibiting the same distribution across all 8 survey sites were excluded from the CCA.


Physical & Chemical

Channel width and depth both demonstrated a positive linear trends, increasing significantly with distance from source (r2= 0.91 & 0.73, p-value <0.001 & 0.007) whilst altitude indicated a significant negative trend (r2= 0.66, p-value <0.001). Discharge however also showed a strong increasing trend with distance downstream (r= 0.87, p-value <0.001).

Water temperature generally increased distance downstream from values of 9.6oC at site 1 to 13.2oC at site 7 however fluctuations occurred between sites 2b & 2a as well as sites 5 & 6 indicting decreased temperatures of 0.5oC & 0.6oC (Fig. 3a). Dissolved oxygen concentrations remained consistent at 100% with a depression of only 1% noted at site 3 and an increase of 2% at site 7 to supersaturated levels. Nutrient concentrations indicated no significant downstream pattern (Fig. 3b) however from site 4 nitrate concentrations appear to fluctuate from high to low values at alternate sites with the highest concentration observed at sites 4 & 5. The highest phosphate concentration was reported at site 2a (0.77 mg L-1) whilst concentrations at site 2b were found to be below level of detection (<LOD).

Hardness also generally increased with distance downstream with values more than doubling from sites 1-7 (24-73 mg L-1) (Fig. 4). Ratios of calcium to magnesium hardness remained similar at sites 1, 2a, 3 and 4 with an average of 1.7:1 however ratio's increased in favour of calcium hardness at all other sites reaching a maximum of 4.3:1 at site 5. Conductivity values mirrored variations in total hardness and whilst pH increased from sites 1 to 2b (6.4 to 7.4), values from sites 3-6 remained close to neutral (7.2) with elevated pH noted at site 7 (7.4).

A combined PCA of both physical and chemical parameters (12 factors in total) resulted in the extraction of three components. All parameters with the exception of nitrate and phosphate concentrations were significant weighted along PCA 1 (>±0.5). Altitude was the only parameter to provide significant weighting to PCA 2 and was therefore ignored due to the significant weighting of altitude within PCA 1. PCA 3 however represented both nitrate & phosphate concentrations (0.56, 0.58) providing an additional dimensions that assessed the nutrient

Figure 3: a) Temperature (oC) & Dissolved Oxygen (%), b) Nitrate (NO3) & Phosphate (PO4) concentrations reported at each of the 8 survey sites

Figure 4: Calcium and magnesium hardness as CaCO3 (mg L-1), conductivity (uS cm-1) and pH values recorded at the 8 survey sites

status of the eight survey sites. It was found that although survey sites could be grouped along PCA 1 into upper, middle and lower sections, additional classification based on nitrogen phosphate ratios (N:P) was also possible. Ratios calculated on transformed (√x+1) concentrations to eliminate the issue values <LOD indicated that groups 1 & 2 (sites 1, 3 & 5) exhibited low average N: P ratios (0.95:1 & 1:1) whilst groups 3 & 4 (sites 2a, 2b, 4, 5 & 6) exhibited higher average N: P ratios (1.2:1 & 1.2:1).

3.2 Biological

Initial observations of macro-invertebrate communities indicate that taxa richness is higher in stream orders greater than 4 (sites 4-7) whereas site 2a, a second order stream, demonstrated the lowest taxa richness with insects representing 90.9% of the total taxa present (Table 2). The contribution of insects to invertebrate communities decreased downstream with site 7 demonstrating the lowest percentage (45.5%). Average score ngfkldnfgndfgmgmdgfmldmgkmdgfkmdgkmdfgkmdfkgmdfkgmfdkmgkdmgkdmgkdmgkdmgkdgmdkgmddglmdlper taxa (ASPT) results indicate however that site 2a demonstrated the lowest ecological status (1.5) whilst sites 4 & 6 scored the highest (2.9).

Table 2: Macro-invertebrate data based on the presence absence of 51 observed taxa (48 families), Average Score Per Taxa (ASPT) generated from the Biological Monitoring Working Party (BMWP) and percentage of total taxa separated by feeding guilds; Scrapers (S), Collectors (C), Grazers, (G) and Predators (P)








Taxa Richness








Total % Insect
























Feeding Guilds


































The division of taxa into feeding guilds indicated little longitudinal shift in dominant feeding strategies with distance downstream (Fig. 6). Predation was the dominant feeding strategy at five of the eight sites (sites 1, 2a, 3, 5 & 7), whilst scrapers with the exception of site 5, remained relatively abundant within all stream orders. Percentage abundance of collectors and grazers within the invertebrate assemblages were more variable with both found in low abundance at site 2a (9.1%) and collectors also only represented 7.1% of total taxa present at site 3. The highest percentage abundance of collectors was identified at site 1 whereas gatherers were most prevalent at site 2b.

3.3 Biological & Environmental Gradients

An examination of species assemblage at an order level indicated that insects as such Ephemeroptera, Tricoptera and Diptera were found to be present at all eight survey sites and Plecoptera absent at site 1 only (Fig. 7). Oligocheates were also found to be present throughout the survey area. Taxa such as Gordiacea and Unionidae were predominately confined to the upper reaches whilst Megaloptera, Hirundinea, and Trichladida were all found to present within the mid-section of the river (orders 3 & 4). Other taxa were found to be present in the upper and lower reaches of the survey area but absent within the middle section (sites 2b-5) and included Hemiptera and Coleoptera. Pulmonates undulated between present and absent with distance downstream whilst Mesogastropods, Decapoda and Odonata were the only orders found exclusively in the lower section of the survey area (sites 5-7).

Figure 7: Presence absence of the twenty-two taxonomic orders identified during surveying



At family level, the distributions of a number of taxa were found to be related to longitudinal environmental gradients (Fig. 8). The direction and length of the lines representing the four environmental variables form biplots with the first synthetic gradient (F1) to indicate the direction and maximum rate of change of environmental conditions (Braak & Verdonschot, 1995), whilst the point of origin (0, 0) represents the average value of each variable. Figure 10 shows that the first synthetic gradient (F1) is negatively correlated to slope and PO4 concentrations but positively correlated to conductivity and total hardness. The proximity of sites and taxa to the environmental endpoints therefore provides an indication of the conditions that can be found at these sites and the associated taxa.

Insect taxa such as Leptoceridae, Nemouridae, Philopotomidae and Baetidae were predominately associated with steeper river gradients resulting in insects representing the highest percentage taxa at upstream sites. Insects such as Nemouridae, Perlidae, Periodida, Leuctridae and Limnephilidae however, were also associated with higher than average PO4 concentrations that were observed at sites 2a & 3. Mollusc and crustacean taxa including Sphaeridae, Hydrobidae, Planorbidae and Gammaridae were found to be present at higher than average conductivity and total hardness levels identified at the lower survey sites (sites 5-7).


The abiotic continuum

A number of abiotic variables measured along the river Tanat demonstrated significant linear trends downstream supporting the underpinning theory of the RCC that environmental gradients occur along the river continuum. The observed changes in physical and chemical parameters were also indicative of temperate rivers. The upper reaches were typically characterised by steep, narrow planforms (Whiting & Bradley, 1993) and shallow flows which resulted in the cooler, well oxygenated lotic environments (Giller & Malmqvist, 1998) observed.

River gradient decreased with distance downstream and as the valley became wider, energy loss to lateral erosion resulted in an increase in channel width (Ritter, 2006). The merging of tributaries increased water volume within the main channel resulting in greater discharge values however increases in velocity were counterbalanced by the decreasing gradient (Ritter, 2009). Temperature decreases downstream are likely to be due to increased isothermal stability of the larger water volume and reduced shading (Allan & Castillo, 2007, p.97).

Total hardness and conductivity demonstrated a significant linear increase with DFS resulting from the accumulation of major ions and calcium carbonate (CaCO3) derived from the merging tributaries (Giller & Malmqvist, 1998, p.53). However, the higher pH and total hardness recorded at site 2b are a result of a volcanic intrusion outcrop which is base rich, highlighting the important effects of geology on water chemistry. This relationship indicates that the river Tanat receives low groundwater inputs and that adjoining tributaries within the upper portion of the river profile have low buffering capacities (Hill & Neal, 1997) due to the draining of Ordovician and Siliurian Shales (Taylor, 2001). The presence of paleozoic mudstones and sanstones further downstream resulting in an increase in buffering capacity, is evident from observed increases in conductivity (Neal et al., 1997).

Fluctuating nutrient concentrations (NO3 & PO4) are likely to be the result of nutrient spiralling (Webster, 1975) which refers to the recycling of nutrients from dissolved inorganic forms, through a particulate and biological uptake phase and their return to the water column in dissolved form following excretion and microbial degradation (Newbold et al., 1981). The spiralling length (S) is dependent on a number of biochemical and geomorphic factors including channel size which in turn controls the surface are to column ratio between the benthos and dissolved inorganic ions as well as the residence time which equates to exposure time (Ensign & Doyle, 2006). Newbold et al., (1982) also highlighted the role of different invertebrates feeding guilds in controlling the regeneration transportation and uptake of nutrients.

Results support the belief that channel morphology provides a useful measure for catchment scale river classification due to the progressive changes in channel structure along the river continuum (Jeffers, 1998). Whiting & Bradley (1993) for example, developed an upland channel classification system based on physical parameters such as gradient, sediment size and channel width and depth. The River Habitat Survey (RHS) also recognises the value of determining physical characteristics of a river to provide an easy to measure assessment of habitat quality (Newson et al., 1998; Raven et al., 2000). However, the idealism that changes in channel geomorphology are unidirectional is considered by Schuum (1997) who although, classified river planforms based on geomorphological processes, noted that in-stream structures are formed by localised conditions and therefore may occur multiple times along the river continua forming a heterogeneous longitudinal profile.

Whilst chemical analysis provides an indication of in-stream processes, it was found that measurements did not significantly enhance habitat classification when coupled with geomorphology due to the covariance of many factors (e.g. temperature, conductivity & hardness) and the non-conservative behaviour of others (e.g. nutrients). In the past physical and chemical parameters have been used an indication of biological status (Storey et al., 1990 & Marchant et al., 1997) however, under new UK initiatives aimed to meet the requirements of the EC Water Framework Directive (WFD) (European Commission, 2000), models such as the River InVertebrate Prediction and Classification System (RIVPACS) are becoming increasing focussed on channel morphology and map derived physical parameters for the classification of macroinvertebrate communities (Dawson et al., 2002). It is therefore evident that at scales greater than reach level, the use of water chemistry as a measure of the river continua introduces a level of scientific complexity that exceeds current river management requirements.

4.2 The Biological continuum

In order correlate changes in environmental parameters and biological assemblages along the river continua, a number of biological classification methods have evolved in an attempt to identify different macroinvertebrate communities. Based on the concept of resource partitioning (Townsend, 1989), Cummins (1973) suggested the classification of invertebrates based on functional feeding groups and this was later integrated into the RCC (Vannote, 1980). The results of this study however, found no discernable pattern between FFG's along the river Tanat and although Hawkins & Sedell (1981) identified shifts in feeding functionality predicted by the RCC, the use of FFG has received considerable criticism primarily due to ability of a number of taxa to alter feeding modes at different life stages and when exposed to different resources (Cummins, 1988; Mihuc, 1997 & Rawer et al., 2000). Taxonomic classification at order level provided significantly greater information regarding community assemblage and with knowledge of dominant feeding strategies, functionality could be inferred at a greater level of detail than the use of FFG's alone.

CCA revealed a number of distinct relationships between abiotic parameters and the presence absence of invertebrate taxa. The greater presence of crustaceans and molluscs at downstream sites was primarily a result of the higher calcium carbonate (CaCO3) levels which is utilised for the assimilation of shells and skeletal structures (Giller, & Malmqvist, 1998, p54). Stoneflies (Plecoptera) are characteristic of colder, cleaner environments (Giller & Malmqvist, 1998, p.87) and therefore a number of taxa where associated with lower order streams. Plecoptera families including Nemouridae and Capniidae are detritivores and their association to sites with high than average PO4 could also indicate sites of significant allochthonous inputs (Vannote et al., 1980). Whilst CCA proved a useful tool in correlating abiotic and biotic parameters, it failed to indicate significant relationship between invertebrate communities and the four environmental variables.

Whilst identifying biological communities associated with specific range of environmental parameters along the river continuum is possible, identifying the forces driving community shifts was found to be considerably more complex. Although the RCC evaluates biological processes that operate on a range of spatial scales for example, the input and utilisation of dissolved organic material (DOM), Vannote et al., (1980) fails to allude to multi-scale biological classifications. The likely reason being, biotic assemblages are altered by so many different environment gradients and disturbances, at a range of scales from community to species level, that it would be a gross over simplification to assume biotic distributions followed a progressive distribution along the river continuum. In could therefore be concluded that the Patch Dynamics Concept proposed by Townsend (1989) provides a more realistic model of abiotic-biotic interactions in lotic systems.


Whilst physical attributes of river systems determined at reach level and above provides significant and already exploited scope for habitat classification, extensive chemical analysis of river systems was found in this case to contribute little to the classification process. However, this is not to say this is the case for all river classifications, and may be significant in the classification of polluted and urban rivers. Classification of taxa based on feeding guilds failed to indicate any significant variation in functional feeding modes along the river continuum and whilst taxonomic assessments at order level provided more information, this too failed to identify discernable differentiation between taxa assemblages. The correlation between significant environmental gradients and presence absence of taxa did however identify a number of the potential driving forces for the distribution of crustaceans, molluscs and insects. Finally, whilst the physical organisation of rivers is often the principle component of reach to catchment scale habitat classification (Dawson et al., 2002), it still remains relatively subjective therefore, synergism with analysis of taxonomic organisation which provides an empirical, more objective procedure will ultimately result in a more accurate classification system.