The effect of water temperature regime on benthic macroinvertebrates
A contribution to the ecological assessment of rivers
Inaugural-Dissertation
zur Erlangung des Doktorgrades Dr. rer. nat. des Fachbereichs Biologie und Geografie an der Universität Duisburg-Essen, CE
vorgelegt von
Alexandra Haidekker
aus Hamburg
Oktober 2004
Die der vorliegenden Arbeit zugrunde liegenden Experimente wurden in der Abteilung Hydrobiologie des Instituts für Ökologie der Universität Duisburg-Essen durchgeführt.
1. Gutachter: PD Dr. Hering (Universität Duisburg-Essen)
2. Gutachter: Prof. Dr. Greven (Universität Düsseldorf)
Vorsitzender des Prüfungsausschusses: Prof. Dr. Schuhmacher (Universität Duisburg-Essen)
Tag der mündlichen Prüfung: 22.02.2005
πάντα ρει̃
Table of Contents
1. General introduction ...... 1
2. The temperature characteristics in two river-types in the Lower Mountain Range of Western Germany ...... 6
2.1. Introduction...... 6
2.2. Materials and methods ...... 7 2.2.1. Study area and study sites ...... 7 2.2.2. Recording of temperature data, assessment of environmental variables ...... 11 2.2.3. Weather data...... 14 2.2.4. Temperature data processing...... 14 2.2.5. Temperature parameters used for the analysis ...... 15 2.2.6. Data analysis ...... 16
2.3. Results...... 18 2.3.1. Thermal parameters – ecological significance...... 18 2.3.2. Thermal parameters – typological significance ...... 21 2.3.3. Small- and mid-sized rivers – are they thermally homogeneous? ...... 23 2.3.4. Thermal pattern in relation to environmental conditions...... 25 2.3.5. Classification with other river-types ...... 27
2.4. Discussion...... 32 2.4.1. Temperature parameters of typological and ecological importance ...... 32 2.4.2. Thermal pattern of the river-types and environmental influence...... 35 2.4.3. Classification with other river-types ...... 38
3. Influence of a modified temperature regime on the development of two Hydropsyche species...... 40
3.1. Introduction...... 40
3.2. Materials and methods ...... 42 3.2.1. Study area and sampling sites ...... 42 3.2.2. Assessment of environmental data...... 43 3.2.3. Invertebrate data: sampling and determination ...... 46 3.2.4. Statistical tests...... 47
3.3. Results...... 48 3.3.1. Thermal changes downstream from the Bigge confluence and the power station...... 48 3.3.2. Comparison of the thermal changes by hypolimnic and cooling water ...... 50 3.3.3. Development of two Hydropsyche spp. in different thermal regimes ...... 52
3.4. Discussion ...... 58 3.4.1. Environmental situation ...... 58 3.4.2. Longitudinal influence of thermal changes...... 59 3.4.3. Development of Hydropsyche larvae ...... 60
4. The influence of water temperature on the macroinvertebrate community...... 65
4.1. Introduction...... 65
4.2. Materials and methods ...... 66 4.2.1. Study area and sites...... 66 4.2.2. Sampling ...... 67 4.2.3. Data processing ...... 69 4.2.4. Data analysis ...... 70
4.3. Results...... 71 4.3.1. Variation of environmental and thermal characteristics ...... 71 4.3.2. The EPTC-community ...... 73 4.3.3. Community pattern and habitat...... 74 4.3.4. Explanatory power of the environmental variables...... 75
4.4. Discussion...... 81 4.4.1. Temperature and community pattern ...... 81 4.4.2. Temperature parameters in comparison ...... 81 4.4.3. Other environmental variables and their effects on community ...... 82 4.4.4. Temperature-sensitive species: do they reflect strategies for different thermal environments?...... 84
5. Summary and conclusion...... 90
6. Zusammenfassung ...... 93
7. References...... 104
8. Appendices ...... 122
1. General introduction
1. General introduction
One early reason for intensifying research in running waters was the rising pollution level in the beginning of the 20th century (Schönborn 2003). Kolkwitz & Marsson (1908, 1909) developed the first saprobic index to assess the organic pollution in rivers. In the following years several other saprobic systems (e.g. Liebmann 1962, Pantle & Buck 1955, Zelinka & Marvan 1961) and biotic indices (e.g. Woodiwiss 1964, Beak 1965) were developed in different regions and countries in Europe in order to assess the water quality. With increasing knowledge and the attention paid to the ecological state of rivers the focus turned also to other stressors apart from organic pollution, since rivers have been modified in many ways: in particular anthropogenic influence caused further changes in stream-water physico-chemistry, i.e. acidification, eutrophication, input of toxic sub- stances, modifications of the thermal regime, as well as changes in river morphology, hydrology, and habitat diversity. An increasing number of assessment systems have been developed to measure the state of degradation for most of the stressors mentioned above, using mainly benthic inverte- brates, but also fish, phytobenthos, phytoplankton, or macrophytes as indicators. The majority of methods has been developed to detect the degree of organic pollution and general degradation in rivers (e.g. Kokes et al. 2001, Schöll & Haybach 2001), followed by assessment methods for hydromorphological degradations (e.g. Buhmann & Hutter 1996, LAWA 2001, Lorenz et al. 2004), eutrophication (e.g. Institut für Wassergüte 2002, Bayerisches Landesamt für Wasserwirtschaft 1998), acidification (e.g. Braukmann 1999, Rutt et al. 1990) and toxic substances (e.g.Wogram & Liess 2001) (for a review on assessment methods see Birk & Hering 2002, Knoben et al. 1995). Recently, the EU Water Framework Directive (WFD; Directive 2000/60/EC – Establish- ing a Framework for Community Action in the Field of Water Policy) sets uniform stan- dards in water management practices for Europe with the focus on biotic indicators to survey the ecological status of the water bodies. For most of the stressors mentioned above experienced assessment methods existed, which were reviewed and partly inte- grated in a new ecological assessment system (Hering et al. 2004). It is remarkable that up to today there are no methods for the evaluation of modifications of the thermal regime and its ecological consequences, although the WFD requires ther- mal parameters as one of the physico-chemical parameters to be assessed for evaluating the ecological quality of running waters (Annex V 1.1 and 1.2.).
1 1. General introduction
Consequently, this leads to the question whether a survey of the thermal regime in rivers is necessary, - which it is, when thermal modifications lead to ecologically relevant de- gradations (e.g. effect the benthic biota).
The factor temperature in rivers The ecological condition of flowing waters depends on five major complex factors: hy- drology, morphology, physical and chemical factors, and biological components. All factors are interrelated to a complex system. Regarding the abiotic environment, flow characteristics and temperature are generally considered to be key factors for organisms in running waters (e.g. Elliott 1978, Humpesch & Elliott 1980, Vannote & Sweeney 1980, Stanford & Ward 1983, Reyiol et al. 2001). The temperature holds a particular position among the abiotic factors, since its influence on an organism may be both direct and indirect: its direct influence on an organism con- cerns, for example, temperature dependent metabolic reactions or enzyme functions. In- direct influences take place via temperature dependent chemical or physical processes in the environment, namely the solubility and effects of substances, e.g. the oxygen content of the water or the toxicity of pesticides. The natural temperature regime in rivers is mainly determined by the following factors: first, water temperature is dependent on the climate, defined by the geographical location (latitude, longitude) and by the altitude, connected with the respective air temperature, precipitation, wind condition, air pressure, humidity and solar radiation. Air temperature, compared to the other factors, is considered to be of major importance concerning its influence on water temperature (Smith & Lavis 1975, Crisp & Howson 1982). Second, hydrologic factors influence the water temperature regime, i.e. the spring, tributaries, groundwater influx and the flow regime. Third, the isolation of the water body by ripar- ian vegetation is of high importance for the temperature regime, because it is attenuating temperature extremes in smaller rivers. Fourth, the river morphology, i.e. the channel form, is important as it determines the exchange of temperature between air and water, and finally, the topography, i.e. the valley shape and orientation is responsible for the daily exposition to solar radiation (Ward 1985). Apart from these factors, the natural water temperature regime in rivers has been modi- fied considerably through anthropogenic influence: for example, severe temperature changes are caused by the deep-water release of impoundments with biologically relevant effects on thermal amplitudes, minima and maxima for several kilometers downstream (e.g. Ward & Stanford 1982, Webb & Walling 1993). In addition to that, the discharge of heated industrial water or cooling water of power stations effects the natural temperature regime, intensifying the rate of metabolism, decreasing dissolved oxygen and increasing the rate of nitrification. Changes in the benthic community were observed due to the arti- ficial rise of temperature (e.g. Descy & Mouyet 1984, Fey 1976, Caspers 1978).
2 1. General introduction
Furthermore, modifications of the groundwater level caused by land-use changes effect the temperature regime in running waters, either directly by changing the amount of di- rect groundwater input, or through changes in the relation of surface runoff to groundwa- ter input, causing modifications in the hydrological regime, which indirectly modifies the temperature characteristics of the river (Hynes 1970, LeBlanc 1997). The removal of riparian vegetation results in less isolation of the water body, therefore leading to extreme temperatures. Besides, higher stream bank erosion caused by the lack of riparian vegetation changes the channel morphology which again effects the thermal regime because of a modified water surface-to-volume proportion (Sweeney 1993, LeBlanc 1997). Climate change is another relevant factor to the thermal regime of rivers, its potential effects being currently discussed (e.g. Gleick 2000, Murdoch et al. 2000).
The influence of temperature on benthic invertebrates The importance of the thermal regime in the evolution, distribution and ecology of the fauna in rivers has long been recognized (e.g. Thienemann 1950, Illies 1969, Hynes 1970, Ward & Stanford 1982). Changes in the distribution and in the community pattern of invertebrate assemblages have been connected to temperature in early works (Ide 1935). Several species have been identified to be restricted in their distribution by the temperature range (e.g. Elliott 1978, Humpesch & Elliott 1980, Reyiol et al. 2001), and species richness is found to be temperature dependent as well (e.g. Vannote et al. 1980, Brussock & Brown 1991, Stanford & Ward 1983, Ward 1982). Furthermore, temperature has been recognized as an important gradient in the physical and ecological continuum of rivers: for example in the concept of a longitudinal zonation of rivers (Illies & Botosane- anu 1963), in the River Continuum Concept (Vannote et al. 1980) or in the Thermal Equilibrium Hypothesis (Vannote & Sweeney 1980). Species distributions are known to change with temperature along elevational and altitudinal gradients (Vannote & Sweeney 1980, Ward 1985, Quinn & Hickey 1990). Moreover, river water temperature is one of the key factors determining the life history characteristics of stream invertebrates. Embryonic development, nymphal growth, emer- gence, metabolism and survivorship of certain taxa are regarded to be temperature de- pendent (e.g. Hynes 1970, Sweeney 1984, Lillehammer et al. 1989, Robinson & Minshall 1998, Watanabe et al. 1999).
Classifications of rivers including temperature Rivers have been classified according to their temperature ranges into different thermal types. Brehm & Ruttner (1926) distinguish between summer-warm and summer-cold brooks, Schwoerbel (1999) classifies rivers according to their maximum temperatures. In their relation between fish regions and temperature ranges, running waters have been
3 1. General introduction
classified into different longitudinal zones (e.g. Illies 1952, 1961, Schmitz 1954). Among other factors varying in altitude and latitude, the thermal gradient is known to be a key factor for the distribution of macroinvertebrates, as discussed in the River Continuum Concept (Vannote et al. 1980), and in the Thermal Equilibrium Hypothesis (Vannote & Sweeney 1980). The basis for thermal classification concepts reach from single tempera- ture measurements to means based on more detailed studies, but the temperature ranges which have been used to distinguish between the different zones and types of rivers were contradictory (Malicky 1978, Moog & Wimmer 1994). It is remarkable that although the importance of temperature for the ecosystem in running waters is undisputed, there is still a lack of detailed, systematic temperature data espe- cially in Central Europe. In the USA several detailed and long-term studies with continu- ous recordings of the temperature regime exist, e.g. in the drainage system of the White Clay Creek, Pennsylvania (Sweeney 1993). Detailed recordings of temperature exist also for New Zealand (e.g. Quinn et al. 1990, 1994, 1997). In Europe, most detailed studies with continuous temperature recording were conducted in several catchment areas of the UK (e.g. Edington 1965, Crisp & Le Cren 1970, Smith & Lavis 1975, Crisp et al. 1982, Crisp & Howson 1982, Webb & Walling 1986), whereas in other European countries similar studies with continuous temperature recordings are scarce, and detailed analysis of the thermal regime of rivers such as the study in the Ardèche in France (Dolédec 1987), in the Elz (Hofius 1971) and in the Dreisam (Bau- meister 2001) in Germany are the exception. Recently, glacial streams and alpine rivers have been thermally characterized in detail (Uehlinger et al. 2003). Until today, the majority of data on thermal conditions is sampled in an unsystematical way; most of the data are scattered and have been collected as background information in the course of ecological site-specific studies. It has repeatedly been criticized that de- tailed studies on the annual water temperature regime with its components including ab- solute minimum and maximum water temperature, annual and daily variation, number of degree days, and seasonal changes have rarely been evaluated simultaneously (Vinson & Hawkins 1998, Malicky 1978, Ward 1985). Fey (1984) noted that sporadic measurements might lead to false estimations of the temperature regime, as they can only offer a far too rough estimation for ecological studies.
The following conclusions can be drawn from the situation described above: The natural water temperature regime represents a key factor of the abiotic environment in rivers. Today, this factor is often modified by anthropogenic influence. Thermal changes were shown to have an effect on benthic invertebrates in various laboratory stud- ies. In spite of these conditions, until now temperature has rarely been evaluated in detail in surveys on the ecological quality in rivers. The guidelines of the WFD stipulate a river assessment based on biotic indicators (macrozoobenthos, fish, aquatic flora), with the ecological status being defined by com-
4 1. General introduction
paring the biotic community composition at a sampling site with the community that is present under near natural reference conditions. This survey method is based on a river- type-specific classification. Thermal data to the currently established typology of rivers meeting the requirements of the WFD (Pottgiesser & Sommerhäuser 2004) do not yet exist. Furthermore, the question as to whether temperature-sensitive taxa could function as indicators for thermal changes in this assessment system has not yet been studied.
Aims of the present study were • to characterize the thermal regime of two river-types with the objective to test if a thermal typology can be established in accord with the typology developed to meet the requirements of the WFD, and if significant temperature parameters for a thermal assessment can be identified, • to analyse a river with an artificially changed temperature regime in detail, i.e. with emphasis on thermal variability, and observe the effects on the development of two benthic species, • to relate variation in the invertebrate community to temperature, to compare the significance of temperature with other environmental factors, and to identify spe- cies related to the thermal environment.
To meet these aims, three main tasks were subject of the underlying study which will be dealt with in three separate chapters as follows:
• the detailed characterization of the temperature regime in two river-types in the Lower Mountain Range of Western Germany, • the influence of a modified temperature regime on the development of two Hy- dropsyche species, • the influence of the temperature regime on the macroinvertebrate community in 20 rivers in the Lower Mountain Range of Western Germany – a field ap- proach.
5 2. Temperature characteristics Introduction
2. The temperature characteristics in two river-types in the Lower Mountain Range of Western Germany
2.1. Introduction
The EU Water Framework Directive (2000) requires thermal parameters as one of the physico-chemical parameters to be assessed for evaluating the ecological quality of run- ning waters (Annex V 1.1. and 1.2.), but still a basis for the evaluation of the thermal re- gime is lacking. The WFD stipulates a stream-typology with corresponding reference conditions in order to assess the ecological quality of the respective rivers. A typology has been established, based primarily on ecoregions (according to Illies 1978a), size classes, altitude classes and the geology of the catchment. Pottgiesser & Sommerhäuser (2004) established and extended this to a biocoenotic relevant river typology. Based on this typology, two stream types have been chosen for the underlying study in order to characterize them thermally. Temperature data of several other stream types have additionally been used for compari- son.
This chapter focuses on 1. reviewing temperature parameters that have shown to be of ecological signifi- cance for stream zoobenthos, 2. examining if the stream types can be separated by a classification based on different temperature parameters and 3. defining, which environmental conditions are most closely associated with the characteristics of this thermal pattern.
6 2. Temperature characteristics Materials and methods
2.2. Materials and methods
2.2.1. Study area and study sites The study sites of both river-types are situated in the ecoregion 9, in the Rheinisches Schiefergebirge1 in Germany, with 11 sites located in the Sauerland and nine sites in the Eifel (Table 2-1). The study area is characterized by siliceous, carbonic rocks. Approximately 50% of the area is forested, the remaining parts are characterized by pasture, agriculture and urban settlements. Both stream types are for the most part located in floodplain valleys, which are predominantly in agricultural use. Forest in the floodplain, if present, is mainly re- stricted to the bank-sides. For a detailed characterization of the thermal regime, 20 representative sites were chosen in two river-types. According to the typology established for the rivers in Germany (Pottgiesser et al. 2004, Pottgiesser & Sommerhäuser 2004) one group of rivers belongs to Type 5, which is formed of small-sized streams (catchment area 10 to100 km²), pre- dominated by coarse-sized gravel. The second group of rivers belongs to Type 9 with mid-sized rivers (catchment area 100 to 1,000 km²), predominated by fine- to coarse- sized gravel. Both groups of rivers are located in the siliceous lower mountains. Study sites for the thermal analysis in these two rivers types were selected with the ob- jective to have similar catchment areas, altitudes, hydrological, morphological and chemical factors within each stream type, whereas vegetation cover and thus shading in the floodplain and catchment area should change from site to site. In addition to the detailed temperature analysis of the two above mentioned stream types, exemplary data of other stream types were used for an extended comparison: external temperature data of larger rivers, of rivers influenced by impoundment as well as tem- perature recordings in the lowlands were included (Table 2-1 continued). All temperature data were integrated in four groups, depending on the stream typology and on the source of the data. The characterization follows the typology applied by AQEM2, which uses ecoregion (Illies 1978), altitude, size of catchment area and geology as descriptors; this was recently integrated in the typology for Germany by Pottgiesser & Sommerhäuser (2004) with emphasis on biocoenotic relevance.
1 Rhenish Slate Mountains 2 The Development and Testing of an Integrated Assessment System for the Ecological Quality of Streams and Rivers throughout Europe using Benthic Macroinvertebrates – Putting the EU Water Framework Di- rective into Practice. Acronym: AQEM; Contract No: EVK1-1999-00132; Assessment of European streams using benthic macroinvertebrates. For details see AQEM consortium (2002).
7
Table 2-1. Study sites for temperature recordings. Group 1: small-sized mountainous rivers, group 2: mid-sized mountainous rivers, group 3: mid- 2. Tem sized low-land rivers, group 4: sampling sites with temperature data acquired externally (various river types). River type according to Pottgiesser & Sommerhäuser (2004), for details see text.
Region: SaLd: Sauerland, Webl: Weserbergland, N.Geest: Lower Geest, S.Geest: Stader Geest, HeBl: Hessisches Bergland, Bl: Bergisches Land. p erature characteristics Federal states: NRW: Nordrhein-Westfalen (Northrhine-Westphalia), H: Hessen (Hesse), RhP: Rheinland-Pfalz (Rhineland-Palatinate), NS: Nieder- sachsen (Lower Saxony), SH: Schleswig-Holstein. CA: catchment area. Recording time: dates of the temperature-registration. * Missing data: Kyll: Sep to Oct/00, Nims: Oct to Dec/00, Eder: Jun to Nov/00, Röhr: Jul to Oct/00.
River Region CA Altitude Recording time (com- Site pro- Group River Site Code type (fed state) [km²] [m.a.s.l.] plete months) tocol 1 Weiße Wehe Hürtgen WWE 5 Eifel (NRW) 11.2 300 Jun/00 - Jun/01 x 1 Kall Lammersdorf KAL (Kao) 5 Eifel (NRW) 17.4 450 Jun/00 - Jun/01 x 1 Erkensruhr Hirschrott ERK 5 Eifel (NRW) 22.5 400 Jun/00 - Jun/01 x
1 Volme Kierspe VOL 5 SaLd (NRW) 17.5 345 Jun/00 - Jun/01 x 1 Waldbach Endorf WAL 5 SaLd (NRW) 9.0 370 Jun/00 - Jun/01 x
1 Röhr Endorf ROE 5 SaLd (NRW) 4.7 370 Jun/00 - Jun/01* x 1 Salwey Obersalwey SAL 5 SaLd (NRW) 15.5 340 Jun/00 - Jun/01 x 1 Elbrighäuser B. Battenberg ELB 5 SaLd (H) 12.6 380 Jun/00 - Jun/01 x 1 Laasphe Bad Laasphe LAA 5 SaLd (NRW) 14.8 375 Jun/00 - Jun/01 x 1 Dreisbach Dreis-Tiefenbach DRE 5 SaLd (NRW) 25.9 270 Jun/00 - Jun/01 x
2 Rur Wiselsley RUR 9 Eifel (NRW) 154.0 360 Jun/00 - Sep/03 x Materialsandm 2 Kyll Densborn KYL 9 Eifel (RhP) 471.8 305 Jun/00 - Sep/03* x 2 Prüm W Waxweiler PRW 9 Eifel (RhP) 286.5 315 Jun/00 - Sep/03 x 2 Our Auel OUR 9 Eifel (RhP) 293.6 360 Jun/00 - Sep/03 x 2 Nims Birtlingen NIM 9 Eifel (RhP) 222.1 245 Jun/00 - Sep/03* x 2 Lenne Altenhundem LEN 9 SaLd (NRW) 190.0 280 Jun/00 - Sep/03 x 2 Nuhne Neukirchen NUH 9 SaLd (H) 134.4 310 Jun/00 - Sep/03 x 2 Eder Röddenau EDE 9 SaLd (H) 523.7 280 Jun/00 - Sep/03* x 2 Orke Reckenberg ORK 9 SaLd (H) 275.0 295 Jun/00 - Sep/03 x e 2 Prüm B Beifels PRB 9 Eifel (RhP) 327.1 280 Jun/00 - Sep/03 x thods
8
Table 2-1. continued 2. Tem
p erature characteristics River Region CA altitude recording time (com- site Group River Site Code type (fed state) [km²] [m.a.s.l.] plete months) protocol 3 Osterau Bad Bramstedt OST 15 N.Geest (SH) 172 13 Jun/02 - Mar/03 3 Schmalfelder Au Bad Bramstedt SFA 15 N.Geest (SH) 300 10 Jun/02 - Mar/03 3 Este Buxtehude EST 15 S.Geest (NS) 300 10 Jun/02 - Mar/03 3 Wümme Scheeßel WUE 15 S.Geest (NS) 288 24 Jun/02 - Mar/03 4 Diemel Helminghausen DIH 9 HeBl (H) 103 337 Jun/00 - Sep/03 4 Diemel Talsperre DIT (9) HeBl (H) 102 340 Jun/00 - Sep/03 4 Eder Schmittlotheim EDS 9 HeBl (H) 1,202 246 Jun/00 - Sep/03 4 Eder Talsperre EDT (9) HeBl (H) 1,443 240 Jun/00 - Sep/03 4 Fulda Guntershausen FUL 10 HeBl (H) 6,366 141 Jun/00 - Sep/03
4 Rhein Köln RHK 10 Bl(NRW) 144,232 34 Jun/00 - Sep/03 4 Rhein Ruhrort RHR 20 Bl(NRW) 152,895 16 Jun/00 - Sep/03 4 Mosel Cochem MOS 10 Eifel (RhP) 27,088 77 Jun/00 - Sep/03
4 Werra Allendorf WER 10 HeBl (H) 5,166 144 Jun/00 - Sep/03
Materialsandm e thods 9
2. Temperature characteristics Materials and methods
The river-types being analysed additionally to the types 5 and 9 mentioned above in- cluded the following: Type 10: large rivers (1,000 to 10,000 km²) in the Lower Mountain Range (ecoregion 9) characterized by gravel; Type 15: mid-sized lowland rivers (ecore- gion 14) mainly with sand and loam; Type 20: very large-sized lowland rivers (ecoregion 14) characterized by sand. In Table 2-1 the sampling sites, location, typology and the source of data are summa- rized.
The characteristics of the four groups of study sites can be summarized as follows: Group 1 (Type 5: small-sized mountainous): ten monitoring sites in ten small-sized rivers in the lower mountainous area of Central Europe: altitude: 280 to 450 m.a.s.l., catchment areas: 13 to 26 km².
Group 2 (Type 9: mid-sized mountainous): ten monitoring sites in nine mid-sized rivers of the same ecoregion: altitude: 270 to 400 m.a.s.l., catchment areas: 143 to 524 km².
Group 3 (Type 15: mid-sized lowland): four temperature monitoring stations were in- stalled in mid-sized lowland rivers. They were located in the Stader Geest in Niedersachsen3 (two monitoring sites, catchment areas: 288 and 300 km²), and in the Niedere Geest in Schleswig-Holstein (two monitoring sites, catchment areas: 172 and 300 km², altitude: 13 and 10 m.a.s.l. respectively).
Group 4 (external data of rivers Type 9, 10, 20): external temperature data has been pro- vided by the Bundesanstalt für Gewässerkunde in Koblenz for nine sites in seven rivers, all located in the same ecoregion as group 1 and 2, i.e. in the lower mountainous region of the Eifel and Sauerland. These monitoring sites include three size types with mid-sized (>100 to 1,000 km²), large-sized (>1,000 to 10,000 km²) and very large-sized catchment areas (>10,000 km²).
3 Lower Saxony
10 2. Temperature characteristics Materials and methods
2.2.2. Recording of temperature data, assessment of environmental variables
Recording of temperature data The temperature of the two river-types was recorded for time periods between one and three years, starting in June 2000. The time periods of temperature recordings at all sam- pling sites are listed in Table 2-1. The temperature was recorded at regular intervals of 30 minutes. For the temperature registration, loggers of the type Gemini Data Loggers Tinytag Plus were used. The log- gers have a resolution of ± 0.2°C. They were calibrated during 24 hours before installing them at the study site. They were also checked and calibrated every time when the data was read out from the loggers about every three months. The loggers were installed at the study site in a way securing that they: • stay under water at all water levels, • are located above the ground to avoid sediments covering the loggers, • are located in flowing, not stagnant water, • are always in the shade to avoid solar radiation heating the loggers directly.
To install the loggers in the stream, steal piles were fixed in the river bottom about 10 to 20 cm above the ground and the loggers were tied to them with rustproof wire. The overall-recording time covered one year for group 1 (small-sized mountainous), three years for group 2 (mid-sized mountainous), ten months for group 3 (mid-sized low- land) and three years for group 4 (external data).
Assessment of environmental variables Each study site of the small- and mid-sized rivers in the mountainous region (groups 1 and 2) was characterized by site protocols including chemical, hydrological and geo- graphical data. These cover characteristics of the catchment area, of upstream and down- stream sections of the sampling site (5 km upstream and downstream from the sampling sites in small-sized rivers, and 10 km upstream and downstream from the sampling sites in mid-sized rivers), and characteristics of the sampling site (250 m upstream and down- stream from the sampling point in small-sized rivers, and 500 m upstream and down- stream from the sampling point in mid-sized rivers (for more details see AQEM consor- tium 2002, Hering et al. 2004). These protocols included more than 130 environmental parameters for each sampling site. For further analysis only those variables were chosen that may possibly effect ther-
11 2. Temperature characteristics Materials and methods
mal characteristics of the rivers (Table 2-2). These were correlated with the temperature data. Catchment size, distance of the study site to the source, altitude and mean slope of the valley floor were taken from topographical maps at scales of 1:25,000 and 1:50,000. River width, depth and current velocity were assessed at the sampling site, the average width was measured from the actual shorelines, mean depth and mean current velocity were calculated from 20 replicates at the sampling site. The amount of forest (in 10 % intervals) in the catchment area and in the floodplain was estimated using the Corine Land Cover System (Statistisches Bundesamt 1997). While in the AQEM site protocol native- and non-native forest is considered separately, the percentages of both forest types were summed up to form one parameter for the analysis in the present study. The parameter “shade” is an average value calculated from the following variables: the left and right shoreline covered with woody riparian vegetation at the sampling site, and the foliage cover shading the channel at the zenith (values in percent). No discharge values were available for the small-sized rivers, but size-related parameters as catchment area and current velocity, width and depth are meant to substitute discharge. In addition to the environmental variables assembled in the AQEM site protocol, the azimuth was measured on topographical maps (1:50,000 and 1:100,000) as the angle (in degrees) at which the overall stream channel differed from due south: i.e. due south = 0°, due west = 90°, due east = -90° . For analysis, the absolute value of azimuth was used, as the amount of radiation a stream receives has to be equal for the orientation of the same magnitude but different sign (Hawkins et al. 1997). As far as environmental variables for river groups 3 and 4 (external data and lowland rivers) are concerned, no site protocols were available for the stations of group 3 and 4; sampling sites with site protocol data are marked in Table 2-1, last column.
.
12
2. Tem
Table 2-2. Geomorphological and environmental variables for the study sites of small- (Type 5) and mid-sized rivers (Type 9). CA:Catchment area, Alt
[masl]: altitude [meters above sea level], Currvel: current velocity (mean), Ca_for: forest in the catchment area, Shade: riparian vegetation 250m (Type5) p erature characteristics /500m (Type9) upstream, Fp_forest: 1km length. Azimuth: channel orientation (for details see materials and methods), Conduct: conductivity. * data not used in analysis, given for later interpretation.** parameter assessed additionally to the AQEM site protocol.
CA Alt Slope Width Depth Currvel Ca_for Fp_for * Shade Azimuth** Conduct* Site Type [km²] [masl] [%] [m] [cm] [m/s] [%] [%] [%] [degrees] [µS/cm] WWE 5 11.2 300 1.4 4.1 18.3 0.52 70 100 90 8 173 KAL 5 17.4 450 2.5 5.0 22.9 0.69 10 80 70 18 185 ERK 5 22.5 400 3.8 3.6 31.6 1.05 70 40 50 16 93 VOL 5 17.5 345 1.1 4.5 25.9 0.61 20 0 10 20 673 WAL 5 9.0 370 1.8 4.5 20.3 0.72 100 70 70 9 163
ROE 5 4.7 370 2.6 3.1 18.7 0.64 90 30 57.5 12 236 SAL 5 15.5 340 1.2 4.0 38.5 1.06 70 0 25 57 265 ELB 5 12.6 380 1.7 4.0 22.0 0.80 100 90 82.5 53 174 LAA 5 14.8 375 2.7 4.0 22.4 0.62 100 80 85 15 150 DRE 5 25.9 270 0.9 4.0 6.9 0.13 90 0 0 50 153 RUR 9 154.0 360 1.0 9 30.5 0.83 40 100 60 60 121
KYL 9 471.8 305 0.4 18 51.8 0.95 50 0 30 5 299 PRW 9 286.5 315 0.5 17 50.5 0.41 40 0 20 15 196
OUR 9 293.6 360 0.2 14 50.8 0.70 40 0 72.5 33 136 Materialsandm NIM 9 222.1 245 0.5 20 23.5 0.59 30 0 85 5 440 LEN 9 190.0 280 0.3 20 53.0 0.91 80 0 0 75 279 NUH 9 134.4 310 0.7 7 43.5 0.86 60 0 10 68 364 EDE 9 523.7 280 0.2 20 48.5 0.96 70 0 42.5 60 208 ORK 9 275.0 295 0.5 15 34.0 0.72 70 0 20 70 328 PRB 9 327.1 280 0.3 18 55.5 0.84 40 50 50 15 204 e thods 13
2. Temperature characteristics Materials and methods
2.2.3. Weather data Daily minimum, maximum and mean air temperature and precipitation, covering the same time period as the temperature registration in the rivers from June 2000 to June 2001, were provided by the weather service DWD (Deutscher Wetterdienst, Offenbach). The weather stations are located within a distance of at most 25 km to the study sites. Concerning the air temperature, the years 2000 and 2001 were on the average warmer compared to long-term records (1961-1990): spring 2000 was warmer by 2°C, the sum- mer by 1 to 1.5°C . The winter 2000/2001 was 1 to 2°C above the long-term records. Precipitation in 2000 to 2001 in the regions of the sampling sites was higher compared to long-term records (1961-1990): Nordrhein-Westfalen4 had 8%, Hessen5 4% and Rheinland-Pfalz6 17% more precipitation, respectively. In the catchment area of the Rhein7, 12% more precipitation fell in the year 2000 and 11% more in 2001 compared to long-term records.
2.2.4. Temperature data processing Water temperature of a one-year period from June 2000 to June 2001 was used for the analysis of small- and mid-sized rivers of group 1 and 2. Missing data due to logger- failure (Table 2-1) were extrapolated using the temperature registration of the following two years; on the basis of these data the relative differences between the temperature of the individual rivers (which stayed fairly constant over three years of temperature regis- tration) was calculated and transferred to the thermal conditions of the year 2000/2001. For the comparison with other river-types (of group 3 and 4: external data and lowland rivers), temperature data of different time periods had to be processed for analysis (Table 2-1). Data of group 2 (mid-sized mountainous) and group 4 (external data) were recorded from June 2000 to September 2003. Those of group 1 (small-sized mountainous) was recorded from June 2000 to June 2001 and of group 3 (mid-sized lowland) from June 2002 to March 2003. To maximize data base within the same time period, temperature recordings of summer 2002 and winter 2002/2003 were used (which also had the advan- tage of avoiding the extrapolation of data). Only the recordings of small-sized rivers (group 1), with temperature data between summer 2000 to summer 2001, did not cover the same time period. Comparison of the monthly means of group 1 (small-sized moun- tainous) and 2 (mid-sized mountainous) of the years 2001, 2002, 2003 with the monthly means of 2000 showed that 2002 was the year with the smallest differences compared to
4 Northrhine-Westphalia 5 Hesse 6 Rhineland-Palatinate 7 Rhine
14 2. Temperature characteristics Materials and methods
2000 (Student t-Test not significant for Jun, Aug, Jan, Mar, May for 2000/2002, while 2000/2001 and 2000/2003 differences were always significant). In all probability this approach implies the smallest error for the small-sized rivers within this time period. The external temperature data set (group 4) was based on daily means, so daily ampli- tudes and temperature extremes were not available. For the analysis of this data, mini- mum and maximum daily means were used instead of absolute minima and maxima, and these values were calculated likewise for the temperature data of groups 1, 2, and 3.
2.2.5. Temperature parameters used for the analysis Temperature parameters were selected to meet two purposes: First, parameters that may be useful for a thermal typology, with the emphasis on the distinction between river- types. This includes testing not only mean parameters as mostly applied in ecological studies in rivers, but also the variability of extreme values or seasonal differences (e.g. winter temperatures or daily amplitudes in spring). Second, parameters that have shown to be of ecological significance for benthic fauna (i.e. benthic invertebrates). The major- ity of ecological studies relate mean temperatures to the biota, but it has been stressed that maximum and minimum temperatures as well as degree-days are of major impor- tance to the fauna (Malicky 1978, Fey 1984, Vinson & Hawkins 1998).
Temperature parameters taken into account in the present analysis can be summarized in four groups: 1. Mean temperature parameters: • means of different time periods (e.g. month, season, year) • degree-days, which are mainly used in ecological studies. Since for the data set in this study mean values and degree-days were highly correlated (r>0.92), only mean values were analysed. 2. Extreme temperature parameters: • absolute minima and maxima recorded within a time period (e.g. month, season, year) • means of minima and maxima for time periods that are ecologically of in- terest, e.g. seasons.
15 2. Temperature characteristics Materials and methods
3. Amplitudes: • yearly amplitude (based on temperature extremes or monthly means) • monthly amplitudes • daily amplitudes (maxima: year, month; mean: year, month) 4. Rates of temperature change: • seasonal warming/cooling rates • daily warming/cooling rates
The rate of temperature change within 24 hours was directly proportional to the magni- tude of change within the same time span, for that reason warming or cooling rates will not be analysed in this study. Since the external data (Group 4) consisted of daily mean values, daily amplitudes and temperature extremes could not be used in the analysis. Consequently, the classification by thermal parameters consisted of the following variables: summer and winter mean temperatures, maximum daily means and minimum daily means measured within the one-year time period, and the yearly amplitude.
Additionally, temperature parameters were assessed that were demonstrated to be of eco- logical significance for benthic invertebrates. These were used for a comparison with parameters defined to be relevant for a typological characterization. For the assessment of these parameters, literature was reviewed to collect information on laboratory experi- ments or field studies that connect development, life-history traits or distribution of in- vertebrates with temperature.
2.2.6. Data analysis Three statistical approaches were used in this study: First, the temperature data (yearly mean, maximum and mean daily amplitude, summer maximum and mean) was tested on a normal distribution using the Shapiro-Wilk W Test. For all temperature parameters the W statistic was not significant (p>0.2), so the hy- pothesis that the data is distributed normally is accepted. The Student t-Test was used to define temperature parameters that differed significantly between the small- and mid- sized rivers of the lower mountainous region. Second, multiple linear regression analysis (forward stepwise regression) was used to extract environmental factors (as listed in Table 2-2) that explained the temperature pat- terns observed in the two data sets.
16 2. Temperature characteristics Materials and methods
Third, a cluster analysis was performed with extreme, mean and amplitude temperature parameters of all data sets to visualize groups of rivers with similar thermal patterns. The method for analysis was minimum variance clustering (Ward’s method) using squared Eucledian Distance. For the Student t-Test and the multiple linear regression analysis Statistica 6.1. StatSoft Inc. (2003) was used. Cluster analysis was performed by PCOrd 4.01, MjM Software (1999).
17 2. Temperature characteristics Results
2.3. Results
2.3.1. Thermal parameters – ecological significance In order to compile temperature parameters that have shown to be of ecological signifi- cance for the benthic fauna, literature has been viewed for studies concerning the influ- ence of different temperature conditions on the life-history traits and on the distribution of benthic invertebrates. Though the majority of these studies are laboratory experiments, which are conducted under constant thermal conditions different from field situations, there also exist several studies on seasonality, extreme temperatures and temperature fluctuations. Some few studies are based on field observations concerning geographic distributions or community changes. The influence of different thermal parameters on the macroinvertebrates are listed in Table 2-3. The importance of mean temperature conditions during specific time periods or seasons has been shown to effect the growth rate and size of larvae and adult invertebrates as well as the emergence. For example, winter temperatures were shown to be important for the growth rate of a spring-emerging Plecopteran species, or the summer temperatures were identified to be important for the emergence timing of an Ephemeropteran species. De- gree-days are correlated with the above mentioned traits, with the egg development, hatching success and the diapause induction and termination. Extreme temperatures, i.e. maxima and minima, effect embryonic development, hatching success, larval development, growth rate and survival. Maximum temperatures influence hatching success and larval survival, while minimum temperatures form thresholds that are required for growth. Maximum and minimum temperatures were also found to be limiting to the distribution of species. Only few studies concern daily temperature fluctuations, the results being contradictory. While some studies found a correlation between magnitude of daily fluctuation and egg development and larval growth rate, others did not observe any connection.
18
Table 2-3. Temperature parameters: typological and ecological values. Typological value: Month and season in which temperature parameters were significantly different for small- and mid-sized rivers with p-values of this study (**Chapter 2.3.2.). Ecological importance: Invertebrate groups, for 2. which a dependence on temperature parameters was found at different life stages. In parenthesis: Numbers of references, see next page. Tem p erature distribution, community Egg: devel, hatching Larvae/nymph: growth Temperature variables typological value** p(Student t-test)** Adult: size, weight phenology - emergence diapause structure success rates, survival Mean temperature values
year yes p<0.05 characteristics Su p<0.001 Spr: (10) Su:(2, 19, 25) Amphi (45), Wi: Ple season spring, summer, autumn Su: Ple (8), Wi: Ple(53) Su: Ephe (58) Spr/Au: p<0.05 Aut:(10, 2) (53), Su: Tri (57) Apr p<0.05, May- Dec, Jan, Feb : Ple (53) Apr: Ple (24) June-Aug: months Apr - Sep Dec, Jan, Feb : Ple (53) Sep p<0.001 Apr: Ple (24) Ephe (58) degree-days /year yes p<0.05 Su: p<0.001 degree-days /season spring, summer, autumn S Ple (46,59), Col/ Ephe (22,40,46,48,55), pr/Au p<0.05 Tri (57) Ephe (58) Ple (59) Apr p<0.05 Dipt:(23) Ple(40) degree-days/month Apr - Sep
May-Sep p<0.001 Extreme temperature values minimum/year no n.s. Ple: (6,13,20,44,49), Mollu (50), Amphip (11) Ephe:(14,26,28) mean min winter no n.s.
maximum/year yes p<0.001 (Jun) comm (12), Tri (30,39), Ple: (6,13,20,49) Ephe,Ple,Tri;Cru,Oli, Dipt (30), Turbell (31) Ephe: (14,26,28) Moll, Crust (47) Tri (56) mean max summer yes p<0.001 Amph (11) Amplitudes* year (based on means) yes p<0.001 year (based on extremes) yes p<0.05 month May, Jun p<0.05 day yearly mean no n.s. Yes: Ple (17,20), Ephe model (Ward & Stanford day max no n.s. (May) (54) Heteropt (52) No: Ephe (54) 1982) Ple (38,59) day monthly mean Nov-May p<0.05 day monthly max Nov, Jan-Apr p<0.05 day season mean autumn, winter, spring p<0.05
day season max winter, spring p<0.05 Results constant temperatures
19 Ephe (27,3,4) Col (1) Ephe (21), Ple Ephe(3), Ple T const Ple(5,7,9,15,16,18,32,33, (5,32,33,34,36) (35)
Table 2-3. continued: list of references 2. Tem
Nr. Reference Nr. Reference Nr. Reference p 1 Aiken (1986) 21 Giberson & Rosenberg (1992) 41 Marten & Zwick (1989) erature characteristics 2 Baumann (1979) 22 Giberson & Rosenberg (1994) 42 Marten (1991) 3 Bohle (1969) 23 Gillooly & Dodson (2000) 43 Milner et al. (2001) 4 Brittain & Campbell (1991) 24 Gregory et al. (2000) 44 Moreira & Peckarsky (1994) 5 Brittain, Lillehammer & Saltveit (1984) 25 Hawkins et al. (1997) 45 Panov & McQueen (1998) 6 Brittain & Lillehammer (1987) 26 Humpesch & Elliott (1980) 46 Pritchard et al. (1996) 7 Brittain (1977) 27 Humpesch (1980) 47 Quinn et al. (1994) 8 Brittain (1983) 28 Humpesch (1982) 48 Rosillon (1988) 9 Brittain (1991) 29 Huryn (1996) 49 Saltveit & Lillehammer (1984) 10 Brittain et al. (2001) 30 Lamberti & Resh (1985) 50 Schöll (2000) 11 Brujis et al. (2001) 31 Lascombe et al. (1975) 51 Snook (2001) 12 Castella et al. (2001) 32 Lillehammer (1985) 52 Sweeney & Schnack (1977)
13 Elliott (1995) 33 Lillehammer (1986) 53 Sweeney & Vannote (1986) 14 Elliott (1978) 34 Lillehammer (1987 a) 54 Sweeney (1978) 15 Elliott (1987b) 35 Lillehammer (1987) 55 Tokeshi (1985) 16 Elliott (1988) 36 Lillehammer (1987b) 56 Wagner (1990) 17 Elliott (1989) 37 Lillehammer et al. (1989) 57 Wagner (2002) 18 Elliott (1992) 38 Lillehammer, Salveit & Brusven (1991) 58 Watanabe et al. (1999)
19 Fairchild & Holomuzki (2002) 39 Lowe & Hauer (1999) 59 Zwick (1996) 20 Frutiger, A. (1996) 40 Markarian (1980)
Results
20
2. Temperature characteristics Results
Table 2-4. Some temperature parameters for small- (Type 5) and mid-sized (Type 9) rivers for the time period June 2000 to June 2001. Cluster numbers see Figure 2-1. Average temperatures: Msu: mean summer temperatures. Mwi: mean winter temperatures. Extreme temperatures: Max: absolute maximum (recorded Jun/2000). Min: absolute minimum (recorded Jan/2001). Daily amplitudes: Adsumx: summer maximum, Adwinx: winter maximum; Adsum: mean summer; Adwinm: mean winter, Adspm: mean spring. Summer: Jun-Aug/2000, winter: Dec/2000-Feb/ 2001, spring: Mar-May/2001. Temperature values in °C.
River Cluster Msu Mwi Max Min Adsumx Adwinx Adsum Adwinm Adspm type number Wwe 5 1 13.0 4.1 18.8 -0.3 3.7 2.8 2.0 1.4 3.1 Kal 5 1 13.0 4.0 18.5 0.6 4.7 2.1 2.5 1.2 3.0 Erk 5 1 12.7 4.4 18.6 -0.1 3.9 2.3 2.1 1.2 2.5 Vol 5 2 14.3 5.5 19.4 1.3 4.3 3.1 2.0 1.3 3.0 Wal 5 1 11.8 4.6 16.9 0.5 4.0 2.5 2.3 1.1 2.5 Roe 5 1 12.9 4.5 18.2 0.1 5.9 2.8 3.0 1.3 2.9 Sal 5 1 12.0 5.2 16.6 1.9 3.8 1.7 2.1 0.8 1.9 Elb 5 1 12.4 3.5 17.2 0.0 4.0 2.2 2.3 1.1 2.4 Laa 5 1 12.3 3.3 17.2 -0.1 3.9 2.4 2.1 1.1 2.9 Dre 5 2 14.4 5.6 20.3 1.3 5.6 2.1 3.0 1.0 2.3 Rur 9 3 14.2 4.1 22.7 0.4 5.6 1.6 2.8 0.9 2.2 Kyl 9 2 15.2 4.6 20.7 2.0 3.8 1.9 2.2 0.9 1.9 PrW 9 3 15.7 4.4 23.1 -0.2 4.7 2.1 2.4 1.0 2.1 Our 9 3 14.8 4.7 23.0 -0.2 5.8 2.0 3.2 1.0 2.4 Nim 9 2 14.1 5.3 18.3 3.2 2.0 1.8 0.9 0.7 1.2 Len 9 2 14.6 4.4 19.5 -0.1 2.9 2.1 1.5 0.9 1.9 Nuh 9 2 14.2 3.7 20.0 -0.2 4.3 2.0 2.3 1.1 2.7 Ede 9 3 16.7 3.8 21.9 -0.2 4.1 2.1 2.6 0.8 1.7 Ork 9 3 15.4 3.3 21.5 -0.1 5.9 2.2 3.1 1.0 2.8 PrB 9 3 16.3 4.3 25.3 -0.2 6.2 2.1 3.4 1.0 2.1
2.3.2. Thermal parameters – typological significance In the typological approach conducted in this study, differences in thermal conditions between the two river-types (small- and mid-sized mountain streams) were tested. The values for mean and extreme temperatures as well as fluctuation parameters are given for each river in Table 2-4. Summer mean temperatures varied between 11.8 and 14.4°C in small-sized rivers, while in mid-sized rivers they ranged between 14.1. and 16.7°C. Small-sized rivers reached maxima of above 20°C; in mid-sized rivers tempera- tures of over 25°C occurred in summer. Winter temperatures were close to 0°C in most rivers, but in some small- and mid-sized rivers they did not fall below 1 or 2°C. In sum- mer, daily amplitudes ranged between 2 and 6.2°C, a maximum value of 7.9°C being recorded in June 2000 in the Prüm at Beifels (PrB: see Appendix 8). To explore the temperature parameters that are significantly different between the two river-types, a Student t-Test was performed on the mean and maximum values of monthly, seasonal and yearly time periods, and daily, monthly, seasonal and yearly fluc- tuations, respectively. The results are listed in Table 2-5: the small-sized rivers were sig-
21 2. Temperature characteristics Results
nificantly colder than the mid-sized rivers in seasonal mean values of spring, summer and autumn, and during the months from April to September.
Table 2-5. Mean of temperature parameters (°C) in small and mid-sized rivers; Abbre- viations: su: summer, au: autumn, wi: winter, sp: spring; p-values for significant differ- ences in thermal characteristics (Student t-Test) are printed in bold letters. Results for degree-days are similar to mean parameters, they are not listed here.
small mid p-value small mid p-value mean values mean mean t-test amplitudes mean mean t-test year 8.6 9.5 0.002 year (mo. means) 9.8 12.6 < 0.001 season su 12.9 15.1 < 0.001 year (extremes) 17.7 21.4 0.001 season au 9.6 10.2 0.028 daily (mean year) 1.8 1.6 0.202 season wi 4.5 4.2 0.484 daily (maximum) 6.2 6.1 0.833 season sp 7.5 8.5 0.002 months daily (mo. mean) Jun 12.9 15.5 < 0.001 Jun 2.8 3.1 0.410 Jul 12.2 14.1 < 0.001 Jul 1.6 1.9 0.323 Aug 13.5 15.8 < 0.001 Aug 2.3 2.3 0.878 Sep 12.2 13.4 0.001 Sep 1.5 1.3 0.482 Oct 9.7 10.1 0.150 Oct 1.2 0.9 0.017 Nov 6.9 7.1 0.420 Nov 1.1 0.7 < 0.001 Dec 5.4 5.3 0.825 Dec 1.0 0.9 0.021 Jan 3.7 3.4 0.307 Jan 1.1 0.8 0.005 Feb 4.3 4.1 0.493 Feb 1.4 1.1 0.004 Mar 5.1 5.2 0.862 Mar 1.7 1.4 0.014 Apr 6.6 7.1 0.018 Apr 2.5 1.8 0.001 May 11.0 13.1 < 0.001 May 3.8 3.1 0.046
daily (mo. max) extreme values Jun 5.2 5.7 0.488 maximum/year 18.1 21.6 < 0.001 Jul 3.5 4.1 0.109 minimum/year 0.5 0.4 0.857 Aug 3.7 3.8 0.919 minimum wi 0.9 0.5 0.380 Sep 2.8 2.7 0.798 maximum su 16.8 19.8 < 0.001 Oct 2.1 1.9 0.328 Nov 2.1 1.8 0.023 Dec 2.2 2.2 0.924 Jan 2.4 1.8 0.039 Feb 2.5 2.0 0.001 Mar 3.4 2.8 0.030 Apr 4.6 3.3 0.001 May 6.0 5.3 0.211
The same applied to the summer extreme values. On the other hand, temperature parame- ters concerning the colder period of the year, i.e. from October to March, were not sig- nificantly different, and winter minimum values were similar in both stream types. Significantly different amplitudes of the two river-types were recorded from November to May, small-sized rivers having higher daily amplitudes in winter than mid-sized rivers. As to the higher amplitudes in spring, the significant difference between the two river-
22 2. Temperature characteristics Results
types was only valid, when amplitudes of the respective months for both stream types were tested in pairs: in mid-sized rivers highest daily fluctuations occurred one month later (in June) than in small-sized rivers (in May). Comparing the mean and maximum daily amplitudes measured during the one-year period independently of the month, the difference was not significant (p=0.833).
2.3.3. Small- and mid-sized rivers – are they thermally homogeneous? In this study, a classification of small- and mid-sized rivers using cluster analysis was performed using the following temperature parameters: summer and winter mean tem- peratures, summer maxima and winter minima, yearly amplitude, and summer and spring maximum amplitudes (Figure 2-1). The cluster analysis revealed three distinct thermal groups of rivers (Table 2-6): the first group with low summer mean temperatures of up to 13°C, with summer maxima below 19°C (max 18.8°C Group 1 Table 2-6), and daily fluctuations in spring and summer be- tween 3.2 and 5.9°C. The yearly amplitude stayed below 11°C. This group consists only of small-sized rivers. The second group shows intermediate summer mean temperatures between 14.1 and 15.2°C (maxima below 21°C). Some of these rivers have higher winter minima (twice 1.3 and once 3.2°C, see Table 2-4), and low temperature fluctuations of only 2°C (Table 2-4). As Figure 2-1 displays, this group consists of two small- and four mid-sized rivers. The third group only differs from the first two groups by having highest summer maximum temperatures between 21.5 and 25.3°C (Table 2-6), (summer mean temperatures range from 14.2 to 16.7°C and are consequently overlapping with the sec- ond group, Table 2-4). The rivers of the third group show a similar range of daily ampli- tudes and likewise low winter minima. This is a group consisting only of mid-sized riv- ers. Thus the cluster analysis revealed that most small- and mid-sized rivers clearly divided up into two distinct thermally homogeneous groups. Another cluster formed with rivers of both types revealed greater similarity within this group than in their corresponding thermal cluster. The two small-sized rivers of this second group, Dreisbach and Volme, were warmer than the remaining small-sized rivers of type 5. Table 2-2 shows that the Dreisbach is shallower and has a lower current velocity. Moreover, it is not shaded at the sampling site and has no forest in the floodplain. The Volme differs from the other small- sized rivers only by its low amount of forest in the catchment area and floodplain and by its lack of shade at the sampling site. The morphological parameters and the current ve- locity are within the range of the other small-sized rivers, but conductivity is high. The four mid-sized rivers of cluster 2 do not differ as clearly as the two small-sized rivers from their corresponding groups concerning environmental variables (Table 2-2).
23 2. Temperature characteristics Results
Cluster analysis for t-parameters Distance (Objective Function) 5.1E-01 7E+01 1.4E+02 2.1E+02 2.8E+02 Information Remaining (%) 100 75 50 25 0 5-W we 5-Kao 5-Erk 1 5-Roe 5-Elb 5-Laa 5-W al 5-Sal 5-Vol 5-Dre 9-Kyl 2 9-Len 9-Nuh 9-Nim 9-Rur 9-Our 3 9-PrB 9-PrW 9-Ede 9-Ork
Figure 2-1. Classification of small- and mid-sized rivers by minimum variance clustering (Ward’s method) using squared Eucledian Distance. For codes and details on sampling sites see Table 2-1. Numbers preceding codes of sampling sites: 5: Mid-sized rivers, 9: Small-sized rivers.
24 2. Temperature characteristics Results
Table 2-6. Mean and range of minimum - maximum of temperature parameters (°C) of the three groups formed by cluster analysis as shown in Figure 2-1.
Summer mean Winter mean Amplitude year Group mean min max mean min max mean min max 1 12.5 11.8 13.0 4.2 3.3 5.2 9.7 7.7 10.9 2 14.5 14.1 15.2 4.8 3.7 5.6 11.2 9.9 12.5 3 15.5 14.2 16.7 4.1 3.3 4.7 13.1 11.7 14.7
Amplitude - max summer Amplitude - max spring Group mean min max mean min max 1 4.2 3.7 5.9 4.5 3.2 5.4 2 3.8 2.0 5.6 4.0 2.2 5.4 3 5.4 4.1 6.2 4.0 3.2 5.3
Maximum Minimum Group mean min max mean min max 1 17.7 16.6 18.8 0.4 -0.3 1.9 2 19.7 18.3 20.7 1.3 -0.2 3.2 3 22.9 21.5 25.3 0.1 -0.2 0.4
2.3.4. Thermal pattern in relation to environmental conditions In order to determine the environmental variables that were most strongly associated with the thermal regime of small- and mid-sized rivers, a multiple stepwise regression was performed. From the site protocol, nine environmental parameters (see Table 2-2) were chosen as independent variables, and the relative influence of these variables on the nine dependent temperature parameters (see Table 2-4) was tested for the data sets of the 20 sampling sites in small- and mid-sized rivers forming one unit, and additional for the small- and mid-sized rivers separately.
The results of the significant models of the multiple stepwise regression analysis are shown in Table 2-7 . Variation in summer mean temperature was related to the size of the catchment area and to current velocity. For small-sized rivers, the amount of forest in the catchment area explained a significant part of the variance in mean summer temperature, additionally to current velocity, whereas only the catchment area was significantly related to summer temperatures in mid-sized rivers.
25 2. Temperature characteristics Results
Table 2-7. Results from forward stepwise multiple regression analysis. Models for winter tem- peratures, for temperature extremes and for amplitudes were not significant. A significant Beta is marked by bold values and an X, Beta* signifying a coefficient for partial correlation of standardized variables. Values for Beta not typed in bold signify non-significant partial correla- tion.
Model: all rivers all rivers all rivers Summer mean Winter mean Amplitude spring F=27.68 F=3.11 F=5.78 p<0.001 p=0.041 p=0.005 R²adj=0.88 Beta* R²adj=0.40 Beta* R²adj=0.05 Beta*
Catchment area X 0.83 -0.50 Altitude 0.32 Slope X -0.91 Width X -0.70 Depth -0.54 0.35 Current velocity X -0.28 0.42 -0.35 Forest in CA -0.18 Shade -0.20 X -0.63 Azimuth 0.13 X -0.85
Model: small-sized small-sized small-sized Summer mean Winter min Summer ampl max
F=10.70 F=11.97 F=7.24 p=0.02 p=0.02 p=0.04 R²adj=0.84 Beta* R²adj=0.86 Beta* R²adj=0.90 Beta*
Catchment area 0.22 -0.37 -0.34 Altitude X 0.63 Slope Width -0.40 X 0.40 -0.38 Depth X -0.88 Current velocity X -0.69 Forest in CA X -0.60 Shade -0.19 X -0.90 X -0.89 Azimuth 0.30
Model: mid-sized mid-sized mid-sized Summer mean Winter min Winter ampl max
F=6.72 F=10.48 F=11.26 p=0.02 p=0.02 p=0.02 R²adj=0.56 Beta* R²adj=0.93 Beta* R²adj=0.85 Beta*
Catchment area X 0.67 -0.28 Altitude Slope X -0.98 Width -0.29 Depth 0.25 X -0.85 Current velocity 0.42 -0.35 Forest in CA 0.74 Shade X -0.54 Azimuth X -1.50 0.19
26 2. Temperature characteristics Results
Winter temperatures in small-sized rivers determined shade as the variable explaining most variance in the model, whereas in mid-sized rivers azimuth was the most important factor related to winter minima, furthermore depth explaining a smaller part of the vari- ability. With regard to the model including all rivers, mean winter temperatures proved to be related to shade, azimuth and slope. Amplitudes in winter and spring were most strongly correlated with stream width. Winter amplitudes in mid-sized rivers were additionally related to depth, slope and shade. Re- garding summer amplitudes, only the model for the maximum values in small-sized riv- ers was significant, extracting depth, shade and altitude as significant parameters. Models of mean temperature values in winter and summer turned out to be often more significant and revealed better correlation coefficients than models of extreme winter- and summer values. As to mean and extreme values of amplitudes, no such difference showed.
2.3.5. Classification with other river-types Another approach to testing similarities of thermal characteristics with more data from other stream types was realized by cluster analysis. It has to be noted that only daily mean temperatures were analysed, so the results of the classification might be different, when daily temperature fluctuations or extreme temperatures are included in the model in the same way as done in the analysis of small- and mid-sized rivers. Sampling sites, and the corresponding thermal parameters are listed in Table 2-8. The result of the classifica- tion of 34 sites is shown in Figure 2-2 .
The sampling sites divide into three main groups, their thermal characteristics are sum- marized in Table 2-9, detailed temperature data to each river in the corresponding clus- ters can be seen in Table 2-8. • The first consists of sites with the lowest yearly amplitudes (7.7 to 10.9°C) caused by low summer temperatures. It is the largest of the three groups, and within this group three homogeneous clusters of rivers emerged: a first cluster (1a in Figure 2-2)- with thermally very similar rivers of low summer- but also low winter temperatures close to 0°C - consists only of small-sized rivers of the lower mountainous region. A second cluster (1b) consists of three different river-types: two small-sized mountainous rivers, one mid-sized mountainous river and two mid-sized lowland rivers, showing low summer but higher winter temperatures (e.g. Nims 3.1°C in Table 2-8) than the rivers above. The third cluster (1c) has summer temperatures of up to 14.9°C (Table 2-9) and higher winter temperatures of 1.3 to 2.3°C (Table 2-8), thus being the group with the
27 2. Temperature characteristics Results
lowest yearly amplitudes of all sites. Two small-sized rivers and a site in a mid- sized river below an impoundment belong to this cluster.
• The second group shows yearly amplitudes between 11.7 and 15.9°C (Table 2- 9), with minimum temperatures around 0°C and mean summer temperatures between 14.2 and 16.7°C. It consists mainly of mid-sized mountainous rivers; two lowland rivers belong to this group as well as a river influenced by a tribu- tary with an impoundment.
• Main characteristics of the third group are high winter temperatures. Three riv- ers are very large-sized rivers with the highest summer temperatures compared to all other data sets (23.8 to 27.4°C in Table 2-8). The remaining three rivers are mid-sized rivers having slightly lower summer temperatures between 22.9 and 27.4°C (Table 2-8).
Small-, mid- and very large-sized rivers could be classified in different groups based on the distinct summer temperatures. On the other hand, the mid-sized lowland rivers do not form a separate group: two rivers are found in a cluster together with two of the warmest small-sized and the coldest mid-sized river. The other two lowland rivers form a group including a cold mid-sized and a large river influenced by hypolimnic water.
28 2. Temperature characteristics Results
Table 2-8. Temperature parameters (°C) for all river types that were used in cluster analysis. Cluster numbers see Figure 2-2. Temperature values: Mean su: summer mean, Mean wi: winter mean, Max dmw: maximum daily mean , Min dmw: minimum daily mean, Ayear: yearly ampli- tude. For codes of rivers see Table 2-1.
River Cluster Catchm Mean su Mean wi Max dmw Min dmw Ayear type number area [km²] Wwe 5 1a 11.2 13.0 4.1 17.1 -0.1 10.5 Kal 5 1a 17.4 13.0 4.0 15.4 1.3 10.5 Erk 5 1a 22.5 12.7 4.4 16.6 0.4 9.6 Vol 5 1b 17.5 14.3 5.5 17.7 2.1 9.9 Wal 5 1c 9.0 11.8 4.6 14.9 1.4 8.2 Roe 5 1a 4.7 12.9 4.5 16.1 0.7 9.6 Sal 5 1c 15.5 12.0 5.2 14.8 2.3 7.7 Elb 5 1a 12.6 12.4 3.5 15.5 0.0 10.7 Laa 5 1a 14.8 12.3 3.3 16.2 -0.1 10.9 Dre 5 1b 25.9 14.4 5.6 18.5 1.8 10.2 Rur 9 2a 154.0 14.2 4.1 19.2 0.9 11.7 Kyl 9 2a 471.8 15.2 4.6 18.7 1.8 12.3 PrW 9 2c 286.5 15.7 4.4 20.8 -0.2 13.0 Our 9 2a 293.6 14.8 4.7 19.7 0.4 11.7 Nim 9 1b 222.1 14.1 5.3 16.8 3.1 10.4 Len 9 2a 190.0 14.6 4.4 19.8 0.0 11.9 Nuh 9 2b 134.4 14.2 3.7 18.1 -0.1 12.5 Ede 9 2c 523.7 16.7 3.8 22.6 0.0 14.7 Ork 9 2c 275.0 15.4 3.3 20.6 -0.1 14.1 PrB 9 2c 327.1 16.3 4.3 21.6 -0.2 13.6 RhKoe 10 3a 144,232.0 23.4 6.8 27.4 3.7 19.0 RhRuh 20 3a 152,895.0 21.3 7.9 25.4 3.0 15.7 Mosel 10 3a 27,088.0 20.2 6.9 23.8 2.9 15.6 Werra 10 2c 5,166.0 16.7 3.4 21.8 -0.8 15.9 DieTs 9 3b 102.0 18.4 3.4 22.9 1.2 17.6 DieHel 9 1c 103.0 9.6 4.2 13.5 1.3 9.6 EdeTs 9 3b 1,443.0 19.7 4.5 23.7 1.3 17.7 EdeSlh 9 3b 1,202.0 17.4 3.3 23.3 2.6 16.1 FuldGun 10 2b 6,366.0 15.4 3.7 17.6 -0.2 13.4 Ost 14 1b 172.0 14.3 4.2 17.2 0.7 10.9 Est 15 1b 300.0 13.7 5.2 17.0 1.3 10.9 Wue 15 2b 288.0 15.8 2.6 18.4 -0.2 14.7 SfAu 14 2b 300.0 15.6 3.5 18.6 0.0 13.0
29 2. Temperature characteristics Results
cluster analysis all rivers Distance (Objective Function) 2E-01 2.4E+02 4.9E+02 7.3E+02 9.7E+02
Information Remaining (%) 100 75 50 25 0
5-Wwe 5-Erk 5-Roe 1a 5-Kao 5-Elb 5-Laa 5-Vol 5-Dre 1b 9-Nim 14-Ost 15-Est 5-Wal 1c 5-Sal 9-DieHel 9-Rur 9-Our 2a 9-Len 9-Kyl 9-Nuh 10-FulGu 2b 14-SfAu 15-Wue 9-PrW 9-PrB 2c 9-Ork 9-Ede 10-Werra 10-RhKoe 3a 20-RhRuh 10-Mosel 9-DieTs 3b 9-EdeTs 9-EdeSlh
Figure 2-2. Classification of all rivers by temperature parameters using extreme, mean and am- plitude values. Minimum variance clustering (Ward’s method) using squared Eucledian Distance. Rivers are coded by river-type number and abbreviation of river names as in Table 2-1.
30 2. Temperature characteristics Results
Table 2-9. Mean and range of temperature parameters (°C) of the respective groups shown in Figure 2-2.
Summer mean Winter mean Group mean min max mean min max 1a 12.7 12.3 13.0 4.0 3.3 4.5 1b 14.2 13.7 14.4 5.1 4.2 5.6 1c 11.1 9.6 12.0 4.7 4.2 5.2 2a 14.8 14.2 15.2 4.6 4.1 4.7 2b 15.3 14.2 15.8 3.4 2.6 3.7 2c 16.2 15.4 16.7 3.8 3.3 4.4 3a 21.6 20.2 23.4 7.2 6.8 7.9 3b 18.5 17.4 19.7 3.7 3.3 4.5
Max daily mean Min daily mean Amplitude year Group mean min max mean min max mean min max 1a 16.1 15.4 17.1 0.4 -0.1 1.3 10.3 9.6 10.9 1b 17.4 16.8 18.5 1.8 0.7 3.1 10.4 9.9 10.9 1c 14.4 13.5 14.9 1.7 -0.1 2.3 8.5 7.7 9.6 2a 19.1 18.7 19.8 0.7 0.0 1.8 12.0 11.7 12.3 2b 18.2 17.6 18.6 -0.1 -0.2 0.0 13.4 12.5 14.7 2c 21.5 20.6 22.6 -0.2 -0.8 0.0 14.3 13.0 15.9 3a 25.5 23.8 27.4 3.2 2.9 3.7 16.7 15.6 19.0 3b 23.3 22.9 23.7 1.7 1.2 2.6 17.1 16.1 17.7
31 2. Temperature characteristics Discussion
2.4. Discussion
2.4.1. Temperature parameters of typological and ecological importance Summer temperatures Taking the results of this study into account, the conclusion can be drawn that summer temperatures differ considerably between the small- and mid-sized rivers: from May to September, monthly mean and maximum temperatures and the corresponding degree- days are discriminating on a highly significant basis. Although the smaller-sized rivers have significantly lower water temperatures during summer than the larger-sized rivers, the individual rivers within one group differ substan- tially in their temperature regime, too. Monthly mean values often hide the fact that ex- treme temperatures may prevail in the water for several hours (see Figure 2-3 for May 2001). To give an example from the present study: two small-sized rivers (Volme and Dreisbach) reached temperatures of 18°C and higher for two to three percent of the time in May 2001, which equals 17 to 22 hours within this time period, whereas on the other hand 16.3°C was the maximum temperature the other rivers of comparable size reached in May. These patterns clearly show that calculating degree-days based on mean values is not sufficient because critical thresholds might be neglected. Similarly, in the works of Crisp & Le Cren (1970) and Crisp et al. (1982) it has been stressed that for many biologi- cal applications mean values as sums of degree days may not be adequate, as the tem- perature may be above a certain baseline for the fauna for several hours a day.
Table 2-10. Comparison of water temperature ranges for the characterization of the biocoenotic zones. Modified after Moog & Wimmer (1994). Macroinvertebrates have been sampled in 2000/2001 in spring and summer at the 20 monitoring sites, and biocoenotic regions have been determined according to the percentage of the commu- nity preferring a certain zone. Ay mean: mean yearly amplitudes.
Ay mean Ay mean Ay mean Illies (1961) DVWK (1984) Hebauer (1986) This study Eucrenal 2 °C Hypocrenal 5 °C Epirhithral 9 °C 8.2 - 10.9 °C < 20 °C 5-10 °C Metarhithral 13 °C 7.7 - 14.1 °C Hyporhithral 8-14 °C 18 °C 12.3 - 14.7 °C Epipotamal 12-18 °C 20 °C Metapotamal > 20 °C 16-20 °C 18 °C Hypopotamal 16-20 °C 15 °C
32 2. Temperature characteristics Discussion
Summer temperatures have been traditionally used for the classification of rivers (e.g. Brehm & Ruttner 1926, Schwoerbel 1999). Illies (1961) established a longitudinal zona- tion of rivers based on the relation of fish regions and temperature zones, a concept, which has been adopted and modified by other authors (e.g. Illies & Botosaneanu 1963, Schindler 1963, Sandrock 1981, DVWK 1984, Hebauer 1986). Most of these concepts are based on sporadic unsystematic measurements, with the consequence that the tem- perature ranges for the differentiation of the biocoenotical zones are contradictory. Moog & Wimmer (1994) analysed these studies in detail, testing the proposals for temperature ranges in rivers with available corresponding biotic data. Values of the DVWK (1984) for monthly summer means and the annual range of minima and maxima by Hebauer (1986) were found to correspond best with the zones based on the biocoenotic data. The temperature ranges of the biocoenotic zones proposed by Hebauer correspond fairly well with the temperature data of this study and the respective zones (Table 2-10). On the other hand, the estimations of the DVWK temperature ranges are lower than in this
study.