Hydrobiologia 516: 69–90, 2004.D. Hering, P.F.M. Verdonschot, O. Moog & L. Sandin (eds), Integrated Assessment of Running Waters in Europe.© 2004 Kluwer Academic Publishers. Printed in the Netherlands.

69

Identification and measure of hydromorphological degradation in CentralEuropean lowland streams

Christian K. FeldInstitute of Ecology, Department of Hydrobiology, University of Essen, Universitätsstraße 5, D-45117 Essen,GermanyTel.: +49(0)201 183-4390. Fax: +49(0)201 183-4442. E-mail: christian.feld@uni-essen.dewww.uni-essen.de/hydrobiology

Key words: assessment, hydromorphology, reference condition, degradation, German Structure Index, lowlandrivers

Abstract

The objective of the current study was to identify hydromorphological variables that are suitable to define anddescribe hydromorphological degradation. Stream type-specific and spatial scale-dependent multivariate analysis(Non-metric Multidimensional Scaling, NMS) of 106 hydromorphological variables derived from 275 samplesat 147 sites and indicator value analysis (IndVal) resulted in the identification of key factors describing hydro-morphological differences in Central European lowland streams. Sample sites represented six European streamtypes from Sweden (1 stream type), The Netherlands (2 stream types), and Germany (3 stream types). The fourlarge-scale hydro(geo)morphological variables: catchment size, geology (‘% moraines’, ‘% alluvial deposits’), andnatural land use (‘% natural forest’) explained inter-stream type differences best. On the smaller site scale, riparianvegetation described inter-stream type differences best.

On catchment scale, ‘% natural forest’, and ‘agricultural land use’ illustrated inter-stream type hydromorpho-logical degradation of all six stream types very well. Four site related variables (‘% wooded riparian vegetation’,‘% shading’, ‘average stream width’, and ‘% macrolithal (cobbles, 20 to 40 cm long) account for hydromorpholo-gical degradation on the smaller reach-scale. An analysis of indicator variables restricted to German stream typesonly resulted in four factors, namely ‘% xylal’ (tree trunks, branches, roots, etc.), ‘no of debris dams >0.3 m3’,‘no of logs >10 cm ∅’, and ‘% fixed banks’ as important descriptors of hydromorphological degradation.Intra-stream type hydromorphological degradation is illustrated for ‘mid-sized sand bottom streams in the Germanlowlands’. For this stream type, a clear gradient of degradation was revealed, and 25 variables were identified toentirely characterize reference conditions and degradation. The variables that described the degradation gradientbest were combined to the new German Structure Index (GSI), which can be implemented to continuously measurehydromorphological degradation.

Introduction

Since the introduction of the European Water Frame-work Directive (WFD) in 2000, physical habitat evalu-ation has a major focus in Europe (Raven et al., 2002).In particular, hydromorphological degradation has be-come an important stressor affecting the instream biotain many Central European stream types (Feld et al.,2002; Lorenz et al., 2004; Ofenböck et al., 2004;Raven et al., 2002). In this context, saprobic indices

have a restricted applicability in stream assessment,since they aim to detect the single stress factor i.e.,organic pollution only. They are not capable of assess-ing other sources of impairment. The WFD, therefore,necessitate the development of new tools to assess theecological quality of streams and rivers (Hering et al.,2004), including hydromorphology. In order to fulfilthe demands of the WFD, stream and river assess-ment must be changed fundamentally from a singleindex system, such as e.g., Saprobic systems (Czech

70

Saprobic Index: CSN, 1998; German ‘Saprobienin-dex’: DEV, 1992; see also Rolauffs et al., 2004) to amore ‘holistic’ approach. The latter refers to multipleindices, capable of assessing the impact of differ-ent habitat pressures on both, the instream biota andthe physical habitat. Running water ecosystems arecontrolled mainly by geological, hydrological, mor-phological, and water chemistry attributes (Franquetet al., 1995; Hildrew, 1996; Richards et al., 1996).The physical habitat controls riverine biota on bothtemporal and spatial scale (Allan et al., 1997; Beiselet al., 1998a, b; Davies et al., 2000 ; Sponselleret al., 2001). In particular, the scale-dependent relationbetween hydromorphology and the macroinvertebratecommunity in streams and rivers has been widely dis-cussed (e.g., Rabeni, 2000; Sponseller et al., 2001;Statzner et al., 2001). Some authors emphasize therole of large-scale variables, such as catchment geo-logy, while others state sub-catchment, such as landuse, and reach-scale habitat attributes, such as riparianbuffer width, to mainly influence instream communit-ies. Moreover, on a finer spatial scale, the influenceof single hydromorphological features, for examplewoody debris or riparian vegetation, on instream biotais well-known and widely discussed (Dudley & An-derson, 1982; Benke et al., 1985; Hoffmann & Hering,2000; Richards et al., 1996).

Several methods to measure habitat quality andhabitat degradation exist (e.g., Agence de l’Eau Rhin-Meuse, 1996 for France; Barbour et al., 1999 for theUSA; LAWA, 2000 for Germany; Raven et al., 1998,2002 for the U.K.). But Raven et al. (2002) have alsoshown that the cited methods lead to different resultsdue to the different definition of ‘near-natural land use’in the French and German protocol. Moreover, the lackof stream type specifity, as is, for example, the case forthe German ‘Strukturgütekartierung’, requires a revi-sion of existing methods to fulfil the demands of theWFD. Due to the complex relationship between hy-dromorphological attributes and instream biota, it stillremains controversial how to define habitat degrada-tion and on what spatial scale(s). Hydromorphologicalassessment within the EU-funded research project‘The development and testing of an integrated assess-ment system for the ecological quality of streams andrivers throughout Europe using benthic macroinver-tebrates’ (AQEM) generally followed the approach tocompare test site characteristics with specific refer-ence characteristics per stream type (Barbour et al.,1999; Raven et al., 2002). Therefore, stream type-specific hydromorphological reference conditions had

to be defined prior to assessment. This step demandsknowledge on the hydromorphological conditions oc-curring under undisturbed conditions (high status)as a basis for the definition of four hydromorpho-logical degradation classes (good, moderate, poor,bad status) as demanded by the five-class classific-ation of the WFD. Three major questions arise: (i)what is physical habitat (hydromorphological) degrad-ation? (ii) which spatial scale is appropriate to describehydromorphological quality? (iii) which groups ofhydromorphological variables (e.g., land use, hydro-graph, mesohabitat, riparian vegetation) are suitedand minimally necessary to measure the impact ofhydromorphological degradation?

In this study, I present results from stream type-specific, as well as scale-dependent, statistical analysisof hydromorphological characteristics of six streamtypes in ecoregions 13 and 14 of Europe (accordingto Illies, 1978). The aim was, (i) to analyse spatialscale-dependent hydromorphological differences, and(ii) to identify hydromorphological variables suited todescribe reference conditions and different states ofdegradation within a single stream type.

Study sites, material and methods

Data collection

In total, 275 samples collected at 147 sites be-longing to six different stream types and distrib-uted over three different countries (Sweden, TheNetherlands and Germany) were analysed (Table 1,Fig. 1). German and Swedish sites were sampledtwice in March/April/May 2000 and June/July 2000,with the exception of sites of stream type D03,which were sampled three times in June and Septem-ber 2000, and March 2001. Dutch sites weresampled once or twice in April/May/June and/or Au-gust/September/October 2000). All sites belong to theCentral European Lowlands (ecoregion 14), exceptDutch sites south of River Rhine, which belong to theWestern European Lowlands (ecoregion 13).

The hydromorphological status of each site wasderived from a set of variables compiled usingthe AQEM site protocol. A detailed descriptionand a downloadable site protocol is available atwww.aqem.de (see also AQEM consortium, 2002;Hering et al., 2003). In total, 130 hydromorpholo-gical and geological variables were recorded on threedifferent spatial scales:

71

Tabl

e1.

Gen

eral

char

acte

rist

ics

ofin

vest

igat

edst

ream

type

s(s

trea

mty

peco

des

acco

rdin

gto

Her

ing

etal

.,20

03).

Stre

amty

peC

ode

Riv

ersy

stem

(s)

Eco

regi

onC

atch

men

tA

ltitu

depH

Con

duct

ivity

No.

No.

Tota

lno.

Tota

lno.

(acc

.to

Illie

s,si

ze(m

a.s.

l.)(µ

Scm

−1)

refe

renc

ere

fere

nce

ofsi

tes

ofsa

mpl

es

1978

)(k

m2)

site

ssa

mpl

es

Smal

lsan

dbo

ttom

stre

ams

D01

Riv

erR

hine

,14

9–15

133

–136

6.7–

8.3

295–

1750

12

1223

inth

eG

erm

anlo

wla

nds

Ijss

el,E

ms

Smal

lorg

anic

type

broo

ksD

02R

iver

Rhi

ne14

0.1–

11.3

30–5

04.

2–7.

420

0–64

04

413

13

inth

eG

erm

anlo

wla

nds

Mid

-siz

edsa

ndbo

ttom

stre

ams

D03

Ijss

el,E

ms,

1412

0–76

025

–60

7.2–

8.5

330–

815

515

1854

inth

eG

erm

anlo

wla

nds

Elb

e,O

dra

(−64

00)a

Mid

-siz

edst

ream

sin

Sout

hS0

5N

orrs

tröm

,14

32–1

005

15–2

005.

2–8.

260

–155

05

1015

30

Swed

ish

low

land

sM

otal

ast

röm

,

Vir

ån,H

elge

å,

Käv

linge

ån,

Saxå

n,R

önne

å,

Lag

an

Smal

lDut

chsl

owru

nnin

gst

ream

sN

01R

iver

Rhi

ne,

13,1

40.

5–19

01–

180

4.4–

8.6

100–

895

3258

7814

1

Meu

se(M

aas)

,

Dre

ntse

A

Smal

lDut

chfa

stru

nnin

gst

ream

sN

02R

iver

Rhi

ne,

13,1

40.

5–13

75–

180

6.5–

8.4

120–

950

68

1114

Meu

se(M

aas)

,

Dre

ntse

A

�53

9714

727

5

a Sin

gle

site

atR

iver

Spre

e(B

rand

enbu

rg,G

erm

any)

.

72

(1) Catchment-related variables consider the wholecatchment from the stream source to the samplesite, for example distance to source, stream or-der, catchment geology, and catchment land use.They were derived from topographical and geolo-gical maps (scale: 1:50 000 to 1:300 000). Whenavailable, land use was measured using ArcViewGIS and data from Corine Landcover (e.g., Stat-istisches Bundesamt, 1997 for Germany). Sincecatchment variables are generally constant over along period of time, they were recorded only oncefor each sample site.

(2) The longitudinal extent of reach-related (up-/downstream) variables depends on the size classof a stream type. For small streams (10–100 km2

catchment area), a stretch of 5 km up- and down-stream of the sample site was taken into con-sideration (= 10 km), whereas in case of mid-sized streams (100–1000 km2 catchment area) astretch of 10 km up- and downstream was analysed(= 20 km). Percent (%) length of impoundments,lack of natural vegetation, or water abstractionrepresent typical up-/downstream variables, whichwere usually derived from topographical maps(scale: 1 : 50 000) and recorded once per samplingsite.

(3) Site-related variables were recorded for eachsampling occasion separately. They refer to astretch of 250 m up- and downstream (= 500 m)of the sample site for small streams and 500 mup- and downstream (= 1000 m) in case of mid-sized streams. Habitat composition and physical-chemical variables are typical site related vari-ables.

Stream characteristics

Sites of ‘mid-sized lowland streams in south Sweden’(type S05) are usually slow-flowing permanentstreams without a distinct valley. The natural low-gradient stream course is usually meandering. Benthicdiatoms represent dominating primary producers inlotic reaches, whereas deep and slow flowing reachesare dominated by macrophytes and epiphytic algae asprimary producers. The prevailing degradation factoris a mixture of organic and nutrient pollution (eutroph-ication), and locally acidification is very important.Degraded sites of this stream type are also hydromor-phologically impaired (e.g., through straightening)and situated in agricultural areas (see also Dahl et al.,2004).

The Dutch streams belong to two stream types:‘Small Dutch slow running streams’ (type N01) and‘small Dutch fast running streams’ (type N02). Thelatter are characterized by higher gradients (meanslope of the thalweg), situated in U-shaped valleyswith higher proportions of gravel on the stream bot-tom. ‘Small Dutch fast running streams’ show apermanent and relatively constant discharge pattern.Stream morphology is always altered by channel reg-ulation and agricultural land use. Thus, high qualityreference sites are almost completely lacking.

‘Small Dutch slow running streams’ (type N01)naturally have a plain floodplain with a meanderingchannel, and instream habitat comprises a higher pro-portion of sand and particulate organic material, whencompared to hill streams. Due to extensive alterationof the stream morphology (straightening, scouring,and removal of floodplain vegetation) and eutrophic-ation, this stream type is almost entirely affected bysevere degradation (see also Vlek et al., 2004).

Pristine (reference) sites of ‘small sand bottomstreams in the German lowlands’ (type D01) are char-acterized by sand of fine to medium grain size and ameandering channel flowing in varying valley forms(trough valley, meander valley, plain floodplain). Or-ganic substrates range from 10 to 50% with a consid-erable amount of CWD (coarse woody debris: logs,debris dams).

‘Small organic type brooks in the German low-lands’ (type D02) are naturally characterized by aU-shaped valley and a braided channel. Organic mi-crohabitats cover most of the stream bottom, forexample phytal [floating stands of Potamogeton poly-gonifolius Pourr. and water mosses such as Sphag-num spp. and Scapania undulate (L.)], xylal (woodydebris, root mats) and CPOM (coarse particulate or-ganic matter: fallen leaves, twigs). The brownish wateris often acidic. Both small stream types have beennearly completely degraded by scouring, straighten-ing, impoundments, stagnation, removal of CWD, anddevastation of floodplain vegetation in the past.

References of ‘mid-sized sand bottom streams inthe German lowlands’ (type D03) are characterizedby sand of fine to coarse grain size, and a sinuateto meandering channel flowing in a meander val-ley or a plain floodplain. Organic substrates coverbetween 10 and 50% of the bottom, of which CWD(logs, debris dams) causes high substrate and currentdiversity. The wide floodplain is dominated by de-ciduous wooded vegetation, and standing water bodies(side arms, backwaters) occur regularly except dur-

73

ing summer when they dry out. Almost all streamsof this stream type have been extensively degradedby scouring, straightening, impoundments, stagnation,removal of CWD, and devastation of floodplain veget-ation due to agricultural land use. Small near-naturalfragments occur in northeastern Germany and Poland(Pauls et al., 2002).

Selection of sampling sites

Due to an extensive sampling programme, the numberof samples taken for a single stream type was re-stricted. Therefore, sample sites were pre-selected ac-cording to a subjective estimation of their degradationstatus. The aim of the pre-selection was a set of sitesthat covered a degradation gradient from reference(high status) to heavily degraded sites (bad status).Degradation was related to the (main) stressor affect-ing a single stream type, which was organic/nutrientpollution (type S05), hydromorphological degradation(types D01, D02 and D03), or general degradation(types N01 and N02). The pre-selection was supportedby information derived from maps, for example, chan-nel form, stream size, stream order or accessibility.Additional information on stream status and streamreaches was compiled using data from earlier stud-ies, monitoring reports, and data on habitat quality,such as the German river habitat survey ‘Struktur-gütekartierung’ (LAWA, 2000). The pre-selection wasthen evaluated during field trips yielding the final setof sample sites.

As a general frame, a set of sites for a singlestream type comprised at least three sites each of asupposed high (reference conditions), good and mod-erate quality, respectively. Poor and bad states wereeach represented by at least one site, so that a min-imum number of eleven sites were sampled per streamtype (see also Hering et al., 2003; Hering et al.,2004). Definition of reference sites followed the ba-sic statements of Hughes (1995) and Wiederholm &Johnson (1996) and aspects defined by Nijboer at al.(2004). When reference sites were not available due todegradation of an entire stream type, the best avail-able sites served as ‘assessment references’, whichwas the case for the Dutch stream types N01 and N02.The ‘assessment references’ represented a ‘good eco-logical quality’ instead of a ‘high ecological quality’according to the WFD.

Evaluation of stream type assignment andhydromorphological degradation

Stream type definition and assignment followed Sys-tem B of the WFD (for detailed description see Heringet al., 2003). When available, stream type tables wereused to support proper stream type assignment (e.g.,LUA NRW, 2001 for German stream types). In addi-tion, hydromorphological variables were analysed tolook for further typologically relevant factors import-ant for proper stream type allocation. The analysis oftypologically relevant hydromorphological variableswas exclusively related to 97 samples of a supposedgood or high quality, since any kind of degradationmay affect or superimpose the results. Six sampleswere excluded from the analysis due to gaps in therespective datasets.

In order to visualize the general structure ofthe environmental dataset, the whole set comprising275 sampling occasions including 106 out of 130 re-corded hydromorphological and geological variables.Twenty-four site protocol variables were excludedfrom the analysis due to the casewise deletion ofmissing data. For the analysis of inter-stream typehydromorphological degradation, a two class classi-fication was introduced, since a reduced classificationwas supposed to facilitate the recognition of a gen-eral hydromorphological gradient. Therefore, samplespre-classified as of high or good hydromorphologicalquality were summarized to the category ‘unstressed’,whereas lower quality sites (moderate, poor or bad)were defined as ‘stressed’.

The hydromorphological degradation of the Ger-man stream types D01, D02, and D03 was analysedusing 90 samples with 104 site protocol variables. Ger-man samples only represented stream types, for whichhydromorphological degradation was the presumedmain stressor.

Development of a Structure Index for mid-sized sandbottom streams in the German lowlands

The German Structure Index (GSI) combines severalstream type-specific hydromorphological features ondifferent spatial scales, such as land use, channel mor-phology, or riparian vegetation, to a single index value.Because the GSI is based on objective variables re-corded from either field surveys or maps, it providesa more objective measure of hydromorphological de-gradation compared to the rather subjective judgmentof the pre-selection. However, the objectivity was in-fluenced in three cases, when weighing factors were

74

Table 2. Hydromorphological variables used to calculate group indices for mid-sized sand bottom streams in theGerman lowlands (D03), with respective spatial scale and calculation formula.

Group Hydromorphological Spatial Calculation

index variable scale formula

‘Positive’ Debris # Debris dams (>0.3 m3), Site 3* # Debris dams + # Logs

Index # Logs (>10 cm diameter)

Organic % Xylal (e.g., dead wood, Site % Xylal/% Organic substrates

substrate branches, roots),

Index # Organic substrates

Shading % Shading at zenith Site % Shading * Average stream

Index (foliage cover), width

Average stream width

Shoreline % Shoreline covered with Reach/ % Shoreline covered with

Index wooded vegetation, site wooded riparian vegetation *

Average width of wooded Average width of wooded

riparian vegetation vegetation

‘Negative’ ‘Positive’/ Presence/absence: Reach/ Backwaters (0/1) – Stagnation

‘Negative’ Index – Backwaters site (0/1) – Straightening (0/1) –

– Stagnation Impoundments (0/1) – Removal

– Straightening of CWD (0/1)

– Impoundments

– Removal of CWD

Land Use % Pasture/grassland Catchment/ % Urban sites * 5 + % Crop land

Index % Crop land reach * 3 + % Pasture/grassland

% Urban sites

Scouring Scouring below floodplain Reach/ Original measure from site

Index level site protocol (cm)

Bank Fixation % Concrete Site % Concrete * 5 + % Stones * 3 +

Index % Stones % Wood/trees

% Wood/trees

used (see below). NMS and subsequently ‘IndVal’analysis (see paragraph ‘Indicator variable analysisusing IndVal’) were used to identify hydromorpholo-gical variables ‘best’-suited to describe hydromorpho-logical degradation. The variables were divided into‘positive’ or ‘negative’, representing either high/goodor moderate/poor/bad hydromorphological conditions.Selected variables were tested for significant differ-ences between the two groups (Mann–Whitney U-test). Redundant variables were identified using cor-relation analysis. However, similar variables may givedifferent information when recorded on different spa-tial scales, and, hence, the information on the hydro-morphological status of a site is also different, even ifstrong inter-correlation between those variables occur.For example a high proportion of native forest in thecatchment indicates the morphological integrity of asite, whereas ‘% shading at zenith (foliage cover)’ of a

site provides information about the riparian vegetationand instream habitat quality itself, without being ne-cessarily linked to a high proportion of native forestsin the catchment. Hence, variables were not automat-ically rejected, if interdependence was high (havinga Pearson’s Correlation Coefficient >0.700). A groupindex was calculated for each variable group, repres-enting a certain habitat quality feature (Table 2). Threegroup indices (‘Debris Index’, ‘Land Use Index’,‘Bank Fixation Index’) were calculated using weigh-ing factors in order to consider the different quality ofcategories present for a single variable. For example,in case of the ‘Bank Fixation Index’, concrete-fixedbanks are weighed higher than stones (rip rap) andstones more than wood-fixed banks (Table 2). ‘Posit-ive’ and ‘negative’ group indices were finally summedup to form the GSI. A list of site protocol variablesused for this study with information on the spatial

75

Figure 1. Location of the 147 investigated sites in Sweden, Germany and The Netherlands. Ecoregion delineation according to Illies (1978),ecoregion numbers in italics.

scale is given in Appendix 1. The GSI was used to cor-relate biota (represented by biocoenotic metrics) withhydromorphological quality of a site (see also Feldet al., 2002; Lorenz et al., 2004; Pauls et al., 2002).

Statistical analysis

Correlation analysis and Mann–Whitney U -tests wereperformed with the XLStat 5.2 statistical soft-ware package (Addinsoft SARL, 2002). The Mann-Whitney-U-Test for non-parametric data was chosen,since frequency plots revealed a lack of normal dis-tribution for all variables. As variables differed innumerical scaling and units of measurement (nom-inal (binary), ordinal, and interval scales), non-metricMultidimensional Scaling (NMS) was used for mul-tivariate analysis, as it provides an appropriate toolfor non-parametric data of different numerical scales(McCune & Mefford, 1999).

To provide comparability between hydromorpho-logical variables of different measurement units, allvariables were standardized by dividing each value bythe square root of the respective variables sum of allsquared values (Formula 1). Thus, the sum of squareswill become 1 for each variable, which equalizes

the contribution of variables to the analysis (Podani,2000).

b = xij

2

√n∑

j=1x2ij

, (1)

b = standardized value xij = raw value of the ithvariable in the jth sample.

All NMS analysis was performed using PC-Ord’s(McCune & Mefford, 1999) ‘autopilot’ settings: afour-dimensional solution as a starting point basedon Bray-Curtis distance measures with medium speedand thoroughness; 15 runs with real data and 30runs with randomized data, and a stability criterionof 0.0001. The variance explained by each multivari-ate axis and Pearson’s Correlation Coefficient forthe correlation of hydromorphological variables witheach multivariate axis were calculated using PC-Ord.Presented two-dimensional ordination plots alwaysshow axes pairs, which explain the maximum vari-ance of the hydromorphological variables used for therespective analysis. The ‘final stress’, a measure thatexplains the discrepancy between the multidimension-ality of the data and the final (low-dimensional) ordin-ation is given. According to Clarke (1993) and Podani

76

(2000), stress values between 0.1 and 0.2 representacceptable results.

Joint plots show the relationship between sampleunits and hydromorphological variables, the latterdrawn as lines radiating from the centroid of the ordin-ation scores. The angle and length of the line tell thedirection and strength of the relationship (McCune &Mefford, 1999). For a given variable, the line formsthe hypotenuse of a right triangle with the two othersides being correlation coefficients (r values) betweenthe variable and the two axes. Only variables (lines)are shown, which r value exceeds 0.500.

‘IndVal’ provides a tool to analyse species as-semblages and uncover indicator species (Dufrêne &Legendre, 1997). In this study, ‘IndVal’ was used ina different way to identify hydromorphological vari-ables that are suited to indicate high or low qualitysites. Therefore, similar to Discriminant Analysis, asite-grouping variable had to be defined prior to ana-lysis. Consequently, results are strongly affected bysubjective judgment on group membership of sites,which was performed during pre-selection of samplingsites. In order to minimize the influence of a subjectivejudgment on statistical analysis and to make group al-location as transparent as possible, NMS analysis wasused a posteriori to determine the number of groupsand the sites belonging to a single group (Fig. 2). Ac-cordingly, the samples were divided into two groups:reference (high status) and heavily degraded (poor orbad status) (Table 3). The two groups represent ex-tremes of the hydromorphological gradient withoutany overlap to adjacent quality classes (Fig. 2) andcomprise 15 samples each. Samples of a pre-classified‘good’ or ‘moderate’ status were omitted.

The better a (hydromorphological) variable ex-plains a group, the higher is the resulting ‘IndVal’index. The highest explanation is reached (i.e., theindex reaches its maximum value of 100 %), if allrecords of a single variable are found in a single groupof samples and if the variable occurs in all samplesof that group. The statistical significance of the ‘Ind-Val’ Index values is evaluated using a randomizationprocedure (Dufrêne & Legendre, 1997).

Results

Stream type assignment

The first two axes of the NMS of the hydromorpho-logical variables account for 83% of its total variance

(Fig. 3). The first axis is correlated mainly with large-scale catchment characteristics, such as catchmentsize, geology, and natural land use practices, whereasthe second axis is correlated with agricultural land useon the catchment scale and the natural shoreline veget-ation and the degree of shading on the reach and sitescale (Table 4). Reach or site-related variables are alsotypologically important, if the substrate compositionat a site is taken into consideration.

Out of the stream types pre-defined using theWFD, five stream types can be identified from Fig. 3:Small organic type brooks in the German lowlands(type D02), small and mid-sized sand bottom streamsin the German lowlands (D01 and D03), and mid-sized streams in the South Swedish lowlands (S05).However, sites of type D01 comprise only two samplesand, thus, lack a sufficient sample size for a valid sep-aration. Taking this into consideration, Fig. 3 revealsonly four stream types. Dutch samples form a dis-tinct cluster separated from other stream types but withconsiderable overlap of Dutch slow running streams(N01) and Dutch fast running streams (N02).

Evaluation of hydromorphological degradation: Allstream types

A gradient of hydromorphologicaldegradation is evid-ent along axis 1 (Fig. 4). Both axes of the NMS plotaccount for nearly 85% of the total variance of theenvironmental dataset. The first axis (60% variance)represents the degradation and is, for example, neg-atively correlated with ‘% land use: native forest’,‘% shoreline covered with wooded vegetation’, and‘% shading at zenith (foliage cover)’ (Table 5). Thesevariables indicate high hydromorphological quality(‘unstressed’) and are represented by sites located onthe left hand side of the NMS plot (empty symbols inFig. 4). In contrast, ‘stressed’ sites are best explainedby, for example, ‘% land use: agriculture’, which ispositively correlated with the first axis of the NMSplot.

The second axis of the NMS ordination plot(Fig. 4) is strongly correlated with catchment geo-logy. Sites dominated by alluvial deposits are situ-ated in the upper part of the NMS plot, whereasmoraine-dominated sites can be found at the bottom.‘(%) land use: native forest’ is negatively correlatedwith NMS axis 2 (Table 5). Sites with a high propor-tion of native forest in their catchment, a rather strongdescriptor of hydromorphological reference condi-tions, are clustered in the lower left corner of the NMS

77

Table 3. Median value and range of hydromorphological variables of stream type D03, signific-antly differing between reference and heavily degraded sites (poor or bad hydromorphological status)(p < 0.001, Mann–Whitney U -test).

Hydromorphological variable Reference Heavily degraded

Median (range) Median (range)

Catchment: Land use: % Native forest 20 (0–40) 0 (0)

Site: Land use: % Native forest 90 (80–100) 0 (0)

Site: Land use: % Total agriculture 0 (0–10) 85 (10–100)

Reach: % Impoundments/dams up-/downstream 0 (0) 85 (40–100)

Site: % Shading at zenith (foliage cover) 80 (60–80) 0 (0)

Site: Average width of wooded riparian vegetation (m) 150 (110–200) 6 (0–16)

Site: # Debris dams (>0.3 m3) 4 (3–22) 0 (0)

Site: # Logs (>10 cm diameter) 63 (35–100) 0 (0)

Site: % Shoreline covered with wooded riparian vegetation 100 (90–100) 20 (0–75)

Site: % Bank fixation stones (rip rap) 0 (0) 100 (20–100)

Site: # Organic substrates 3 (2–5) 1 (0–2)

Site: Max. current velocity (cm s−1) 43 (31–63) 26 (7–53)

Table 4. Pearson’s Correlation Coefficient of hydromorphological variables with the first two NMS axes of the ordination oftypological aspects (Fig. 3). Only correlations >0.500 listed.

Axis 1 r Axis 2 r

Catchment: Geology: % Moraines −0.884 Site: % CPOM 0.608

Catchment: Land use: % Native forest −0.851 Catchment: Land use: % Pasture −0.585

Catchment: Geology: % Alluvial deposits 0.823 Catchment: Land use: % Agriculture −0.559

Catchment: Land use: % Wetland −0.678 Site: % Shading at zenith (foliage cover) 0.524

Catchment: Land use: % Non-native forest 0.608 Site: % Shoreline covered with wooded vegetation 0.507

Site: % Psammal/psammopelal (sand/sand and mud) 0.596

Site: Average stream width −0.596

Site: % Macrolithal (cobbles, 20–40 cm long) −0.585

Catchment: Distance to source −0.567

Catchment: Catchment area −0.566

Site: % Megalithal (cobbles and blocks >40 cm) −0.555

Site: % Shoreline covered with wooded vegetation −0.531

Catchment: Geology: % Acid silicate rocks −0.526

Reach: Altitude −0.523

Catchment: Geology: % Organic formations −0.519

Table 5. Pearson’s Correlation Coefficient of hydromorphological variables with the two NMS axes of the ordination of habitatdegradation (Fig. 4). Only correlations >0.500 listed

Axis 1 r Axis 2 r

Catchment: Land use: % Native forest −0.713 Catchment: Geology: % Moraines −0.763

Site: % Shading at zenith (foliage cover) −0.630 Catchment: Land use: % Native forest −0.727

Site: % Shoreline covered with wooded vegetation −0.595 Catchment: Geology: % Alluvial deposits 0.662

Catchment: Land use: % Wetland −0.506 Catchment: Land use: % Non-native forest 0.573

Catchment: Land use: % Agriculture 0.509 Site: Average stream width −0.559

Site: % Macrolithal (cobbles, 20–40 cm long) −0.525

78

Figure 2. NMS joint plot of 95 hydromorphological variables of 54 samples of ‘mid-sized sand bottom streams in the German lowlands’.Lines indicate variables suited to describe the hydromorphological status best (cut-off level: 0.500), and arrow shows the gradient ofhydromorphological degradation. Final stress: 0.114.

Table 6. Pearson’s Correlation Coefficient of hydromorphological variables with NMS axes of the ordination of habitat degradationin German stream types (Fig. 5). Only correlations >0.500 listed.

Axis 1 r Axis 2 r

Site: % Xylal (e.g., dead wood, branches, roots) −0.761 Catchment: Geology: % Alluvial deposits −0.651

Site: % Shading at zenith (foliage cover) −0.750 Catchment: Land use: % Open grassland/bush land 0.637

Site: % Unfixed banks −0.725 Site: # Logs (>10 cm diameter) 0.594

Site: # Logs (>10 cm diameter) −0.700 Catchment: Geology: % Sander 0.505

Site: % Bank fixation stones (rip rap) 0.666 Catchment: Geology: % Moraines 0.502

Site: % Shoreline covered with wooded vegetation −0.657

Catchment: Land use: % Urban sites 0.612

Reach: % Impoundments 0.600

Site: # Organic substrates −0.576

Site: % CPOM −0.569

Site: # Debris dams (>0.3 m3) −0.537

Catchment: Land use: % Native forest −0.536

79

Figure 3. NMS ordination plot of 97 reference samples of six European stream types (see Table 1). Final stress: 0.155.

Figure 4. NMS ordination plot of 275 samples of six investigated stream types (explanation of stream types in Table 1). Symbols indicate streamtype and status of degradation pre-classified as ‘U’ = unstressed (empty symbols, pre-classified ‘high’ or ‘good status’) and ‘S’ = stressed (filledsymbols, pre-classified moderate, poor, or bad status). Final stress: 0.172.

80

Figure 5. NMS joint plot of hydromorphological degradation of 90 samples of three German stream types (D01, D02, and D03). Lines indicatevariables that describe high and low quality sites best (cut-off level: 0.500). Arrows indicate gradients of hydromorphological degradation.Final stress: 0.108.

plot (in particular stream type S05). Fig. 4 revealsa clear gradient of hydromorphological degradationfor the German stream types D01, D02, and D03(see also Fig. 5), coinciding with the presumed mainstressor ‘hydromorphological degradation’ for thesestream types. In contrast, stream types S05, N01, andN02 show a considerable overlap of ‘unstressed’ and‘stressed’ sites.

Evaluation of hydromorphological degradation:German stream types

A gradient of hydromorphological degradation is evid-ent along axis 1 (Fig. 5) for both, small and mid-sizedstreams. This gradient is best explained by site-scaledvariables (Table 6). In particular, the proportion andnumber of organic substrates on the stream bed, theproportions of wooded shoreline and bank fixation ex-plain the gradient due to their correlations with axis 1.On a catchment scale, it is the proportion of urbanareas that indicates hydromorphological degradationfor the three German stream types. The separationof small and mid-sized samples along axis 2 is pre-dominantly based on catchment geology (‘% alluvialdeposits’ vs. ‘% moraines’), ‘% land use: grassland’,and ‘# logs > 10 cm ∅’ on the stream bed (Table 6),the latter being more frequent in mid-sized streams.However, the pre-classified hydromorphological refer-

ence site of ‘small sand bottom streams in the Germanlowlands’ (D01) is clustered with the reference sitesof mid-sized sand bottom streams (D03). High sharesof organic substrates characterize the respective site(D01 0001 in Fig. 5). In particular, ‘# logs > 10 cm ∅’on the streambed and stream width resembled thoserecorded for D03 reference sites.

Hydromorphological degradation of type D03 canbe derived almost entirely from the site protocol vari-ables, as reflected by a clear gradient for this streamtype. The overlap at the transition from good tomoderate and from moderate to poor status (Fig. 5)disappeared, when stream type D03 was analysed sep-arately (Fig. 2). Here, the pre-classification is wellreflected by the NMS ordination, which accounts foralmost 88% of the total variance in the environmentaldataset. A similar result is evident for ‘small sandbottom streams in the German lowlands’ (D01) and‘organic type brooks in the German lowlands’ (D02),when analysed separately (not shown here). Hence,the three German stream types, as well as their hy-dromorphological status, can be identified solely byenvironmental parameters recorded in the site pro-tocol.

81

Table 7. ‘IndVal’ results of suitable core variables to describe the hydromorphological status of a sample of stream type D03 (significance level:<0.05, based on random samples and 499 iterations). ‘Positive’ variables indicate reference conditions (high quality), ‘Negative’ variables heavilydegraded conditions (poor or bad quality). (IV = ‘IndVal’ index).

‘Positive’ variable IV ‘Negative’ variable IV

Site: Max. current velocity (cm s−1) 95.54 Reach: Land use: % Urban sites 100.00

Site: # Logs (>10 cm diameter) 75.63 Reach: Culverting up-/downstream 100.00

Reach: Land use: % Native forest 63.52 Reach: # Dams obstructing migration up-/downstream 100.00

Site: Average width of wooded riparian vegetation 61.62 Site: % Bank fixation stones (rip rap) 56.23

Catchment: Land use: % Native forest 60.31 Site: % Bed fixation stones 50.00

Site: % Xylal (e.g., dead wood, branches, roots) 55.56 Reach: % Impoundments/dams 45.07

Site: # Debris dams (>0.3 m3) 50.65 Reach: # Transverse structures (e.g., weirs, dams, bridges) 44.69

Site: % Unfixed banks 43.10 Reach: Stagnation 43.80

Site: % CPOM 35.46 Reach: Straightening 38.46

Site: % Shoreline covered with wooded riparian vegetation 33.01 Site: Removal of coarse woody debris (CWD) 30.30

Site: CV depth 27.50 Reach: Channel form 27.95

Site: % Shading at zenith (foliage cover) 43.48 Site: Scouring 25.00

Site: # Organic substrates 29.90

Figure 6. Correlation of ‘% native forests in the floodplain’ andinstream ‘number of logs’ for 12 sites of stream type D03.

Development of a Structure Index for ’mid-sized sandbottom streams in the German lowlands’

In total ‘IndVal’ analysis revealed 25 variables, whichsignificantly describe the end points of the hydromor-phological gradient (Table 7). The variables can beseparated into those, which predominantly indicatereference conditions (‘positive’) and those which areconnected with a heavily degraded hydromorphology(‘negative’). Some variables revealed a considerablecorrelation, as it was for example evident for the pro-portion of native forests on catchment and reach scaleand the number of logs in the stream channel (Fig. 6).

Measures of several hydromorphological variableswere significantly different between reference and

heavily degraded sites (Table 3). Consequently, heav-ily degraded sites are mainly characterized by extens-ive agricultural land use in the floodplain, extensivebank modification, lack of dense riparian wooded ve-getation, and thus lack of shading of the channel andwoody debris on the stream bottom. In addition, onlya small amount of organic substrate occurs at sites of apoor or bad hydromorphological status, and hydrologyis strongly affected by stagnation due to weirs, whichreduce maximum current velocities significantly.

In a next step, variables representing a certainhabitat quality feature (e.g., woody debris, channelmodification, or land use), are combined to group in-dices. Group indices are related to different spatialscales. Altogether, eight group indices were definedand calculated (Table 2).

The ‘Debris Index’ weighs debris dams more(factor 3) than logs, for debris dams provide a higherhabitat complexity and diversity. The relative ‘% xy-lal’ in relation to the ‘total % organic substrates’ isamalgamated to the ‘Organic Substrate Index’. As themaximum degree of shading usually decreases withincreasing stream channel width, the ‘Shading Index’considers both by the relation to the width-dependentmaximum value. However, if a sample site is nearlycomplete shaded, 100% is taken as the resulting shad-ing index independent of the respective stream width.The ’Shoreline Index’ refers to the two dimensionalextension of the wooded riparian vegetation (along thestream course as well as in the floodplain), and thusassesses the buffer strip functionality. Certain ‘posit-

82

Figure 7. German Structure Indices (GSI) for 54 samples of ‘mid-sized sand bottom streams in the German lowlands’ (D03) in decreasingorder.

ive’ and ‘negative’ hydromorphological features are –on a presence/absence level – combined to the ‘Pos-itive/Negative Index’. The extent of land use in thefloodplain is summarized with the ‘Land Use Index’,and a weighing factor allows for the severity (urbanareas > crop land > pasture, meadow or open grass-land). The ‘Scouring Index’ directly represents themeasured incision depth of the stream channel. The‘Bank Fixation Index’ is related to the total share offixed banks, and different qualities of fixation are al-lowed for by weighing (concrete > stones > wood ortrees). For each sample, group indices are calculatedand related to the respective stream type-specific max-imum value of a single index plus 10%. Thus, eachindex value is related to a 110% basis, which con-siders that the samples do not necessarily reflect thebest (or worst, respectively) conditions present for acertain stream type. An addition of 10% was supposedto be sufficient, since reference sites of stream typeD03 already represent a relatively high hydromorpho-logical quality. Afterwards, re-scaled percent valuesof ‘negative’ group indices are simply added up andsubtracted from the sum of ‘positive’ group indices.The resulting value represents the German StructureIndex (GSI). Results for 54 samples of mid-sized sand

bottom streams in the German lowlands are presented(Fig. 7).

Discussion

The objective of this study was to identify suitablevariables to describe hydromorphological degradationof stream types in ecoregions 13 and 14 of CentralEurope. If data analysis was changed from severalstream types to single stream types only, the respectivescale of hydromorphological variables also changedfrom catchment scale to reach or site scale. Thus,the set of hydromorphological variables to identifyhydromorphological degradation strongly depends onthe spatial scale. Earlier studies have also stressedthe role of spatial scale in physical habitat assessment(Richards et al., 1996; Allan et al., 1997; Davies et al.,2000; Sponseller et al., 2001), and some have argueda distinct spatial hierarchy exists that influence envir-onmental variables in riverine habitats (Frissell et al.,1986; Rabeni, 2000). The results of this study supportthis hierarchical organisation of hydromorphologicalvariables.

83

Stream type assignment

According to ‘System A’ of the WFD, ‘surface waterbody types’ can be characterized by four factors: eco-region (according to Illies, 1978), altitude, catchmentsize class, and geology. Those factors usually referto a relatively large area and reflect the common useof spatially large scaled variables for the analysis oftypological aspects (e.g., Omernik, 1987 and Whittieret al., 1988 for the U.S.A.; EU commission, 2000for Europe; LUA NRW, 2002 for the Federal Stateof North Rhine-Westphalia, Germany). In contrast to‘System A’, ‘System B’ considers several obligatory(e.g., altitude, latitude, longitude, geology) and addi-tional variables (e.g., distance to source, mean depth,valley shape, substratum composition). The results ofmy study support the typological relevance of thesehydrological and geological variables. Catchment geo-logy, altitude, substrate composition, and stream andcatchment size are clearly suitable to discriminatebetween investigated stream types of ecoregions 13and 14 in Central Europe (Table 4). In addition, thecurrent study revealed land use characteristics as im-portant typological variables on catchment scale. Forexample, the ‘% native forest’ correlates very wellwith axis 1 of the typological NMS (Table 4). How-ever, catchment land use characteristics reflect thedegree of human activities in the catchment, and,thus already reveal hydromorphological degradation.In case of type S05, both outlier samples (Fig. 3) wereinfluenced by high shares of agricultural land use andtherefore likely do not represent real hydromorpholo-gical reference sites. Allan et al. (1997) and Richardset al. (1996) found catchment geology and land useattributes, in particular the proportion of row-crop ag-riculture, to be strong descriptors of stream habitatconditions and macroinvertebrate communities. Theland use-controlled discrimination between lowlandstream types of Central and Western Europe in thisstudy does not correspond very well with the potentialnatural vegetation expected for this region. The naturalvegetation of both ecoregions is deciduous forest (El-lenberg, 1996). Land use appears to reflect degrada-tion rather than typological aspects. The considerationof additional site scale hydromorphological features,such as ‘% shoreline covered with wooded vegetation’and ‘% shading at zenith’ supports this assumption.Both variables are closely related to degradation, anddense riparian vegetation, usually dominated by Alnusglutinosa (Black Alder) and Salix spp. (Willow), canbe expected along streams and rivers in ecoregions 13

and 14 (Ellenberg, 1996). In regard to catchment landuse properties, the reference dataset considered forthis study does not appear to fulfil the essential re-quirements on reference conditions (Hughes, 1995;Wiederholm & Johnson, 1996; Hering et al., 2004).

The Dutch stream types N01 and N02 were notseparated when using hydromorphological variableson a large spatial scale (Fig. 3). It seems that theyare similar from a hydromorphological point of view,which is also the case for the small German streamtypes D01 and D02. In case of N01 and N02, thismakes sense, since the pre-selection of the Dutchsites was not focussed on the detection of hydromor-phological degradation. Moreover, this is a matter ofspatial scale chosen in the study, and stream type dis-crimination presumably becomes clearer, when ana-lysed on smaller spatial scales, for example, on asub-catchment or reach scale.

Evaluation of hydromorphological degradation

The analysis of the hydromorphological degradationreveals two groups. The first group comprises Dutchtypes N01 and N02 but also the Swedish type S05. Thesecond group consists of the German types D01, D02and D03. Hydromorphological degradation was de-tectable for German types and samples, whereas Dutchand Swedish samples of various pre-classified qualityclustered together (Fig. 4). This reflects the fact thathydromorphological degradation was the presumedmain stressor only for German stream types. Thus,it is not surprising that German sites were orderedalong a hydromorphological gradient and Swedish andDutch sites were not. The presumed main stressorfor the Swedish stream type was nutrient pollution,whereas general degradation was presumed to mainlyaffect Dutch stream types. Swedish samples cluster onthe opposite site of the ordination space compared toDutch samples (Fig. 4). Consequently, Swedish sitesare only weakly affected by hydromorphological de-gradation, whereas Dutch sites are predominantly inmoderate to bad hydromorphological condition. Thisis evident by comparing, for example, the land usecategory ‘% natural forest’, which is zero in case of allDutch samples and ranges from 20–90% (mean: 63%)for Swedish samples. Consequently, hydromorpholo-gical degradation strongly affects the Dutch streamtypes.

The analysis of hydromorphological variables onstream type scale was mainly governed by catchmentproperties, of which only land use characteristics re-

84

flect the degree of human impact. However, on reach-and site-scale, several variables, such as ‘% shorelinecovered with wooded vegetation’ and ‘% shadingat zenith’, were shown to be suitable descriptorsof hydromorphological impact. Therefore, environ-mental variables, compiled to evaluate the physicalhabitat quality, should include small-scale variablesmeasured for stretches of 10 up to 1000 m. TheAQEM site protocol considers different spatial scales,of which only catchment properties and some up-/downstream (stretch of 500–1000 m) variables areavailable through topographical and geological maps.Thus, physical habitat assessment necessitates fieldwork to obtain several important small-scale vari-ables. The role of small-scale hydromorphologicalvariables becomes evident by restricting the analysisto German stream types. Here, small-scale variablesare major descriptors of hydromorphological degrad-ation, in particular the amount and quality of organicsubstrates (woody debris, CPOM) and variables de-scribing riparian vegetation and channel modification.Urbanization and ‘% native forest’ are subordinatehydromorphological features on catchment scale asindicated by lower r-values in Table 6. However,Jones & Clark (1987) and Benke et al. (1981) stressedthe role of urbanization as a major impact on thebenthic invertebrate communities.

There is a clear hydromorphological gradient atsmall and mid-sized sand bottom streams, as well asfor small organic type brooks in the German low-lands (Fig. 5). The different stream types can bedescribed by similar habitat attributes. In particular,woody debris appears to be an important factor influ-encing the hydromorphological status of these streamtypes (Harmon et al., 1986; Gurnell et al., 1995; Her-ing & Reich, 1997; Mutz, 2000). Riparian buffer stripsare important to control the influence of sediment in-put from row-crop agricultural areas on the riverinebenthic community (Newbold et al., 1980; Allan et al.,1997; Tabacchi et al., 1998). Newbold et al. (1980)defined a minimum width of 30 m for riparian bufferstrips as sufficient to provide optimal habitat con-ditions for macroinvertebrates. Allan et al. (1997)stressed the role of riparian buffer strips as a barrier fornutrient supply and sediment delivery. The importanceof both, a dense and wide riparian buffer is also madeevident in the current study. The ‘IndVal’ analysis ofhydromorphological variables for type D03 (Table 7)revealed the ‘% shoreline covered with wooded veget-ation’ and the ‘average width of wooded riparian ve-getation’ to significantly differ between reference sites

and sites of a poor to bad hydromorphological status.Reference sites of ‘mid-sized sand bottom streams inthe German lowlands’ were characterized by ripariantrees, which covered 90–100% of the shoreline andextended between 110 and 200 m into the floodplain.It appears that the extent of riparian vegetation in thefloodplain plays a major role, which is accounted forin the calculation of the ‘Shoreline Index’ (Table 2).

The separation of German lowland stream typeswas, amongst other variables, controlled by catch-ment geology. Geology differed between catchmentsof mid-sized and small streams, however, this is crit-ical when applied to the entire Central lowlands ofGermany. The Central lowlands of Germany can be di-vided by the borderline of the last (‘Weichsel’) glacialperiod. The majority of mid-sized sites of the currentstudy were located in East Germany, which is domin-ated by moraine and sander deposits of the ‘Weichsel’glaciers. In contrast, all small sites were located inthe part of West Germany that was unaffected by the‘Weichsel’ glaciers. This area of the West Germanlowlands is generally dominated by alluvial (fluvi-atile) deposits (Bundesanstalt für Geowissenschaftenund Rohstoffe, 1993).

‘Mid-sized sand bottom streams in the Germanlowlands’ (type D03) clearly clustered apart fromsmall streams (types D01 and D02) (Fig. 5), even ifsites are in a poor to bad hydromorphological status.This underlines the classification as an own streamtype. The subjective pre-classification of sites of thisstream type was reflected by the more objective field-recorded and map-derived variables. Thus, streamtype D03 allowed the definition of distinct hydro-morphological degradation classes using AQEM siteprotocol variables (Fig. 2). Even if a separation of thetwo small types D01 and D02 was not possible, whenanalysed together with stream type D03 (Fig. 5) a sep-arate analysis of the small types (not included in thispaper) showed that both types can be separated solelyfrom the hydromorphological variables recorded in theAQEM site protocol.

On a regional scale (level of one stream type),hydromorphological degradation appears to be bet-ter described by site scale variables (Table 3). Thus,site related physical habitat evaluation is especiallyimportant, when habitat evaluation is applied on asmaller spatial scale. Several methods integrate thissite related evaluation in Europe, such as the BritishRiver Habitat Survey (RHS, Raven et al., 1997, 1998,2002), the German ‘Strukturgütekartierung’ (LAWA,2000) or the French SEQ-MP (Agence de l’Eau Rhin-

85

Meuse, 1996). However, these methods do not coverall variables listed in Tables 3 and 7. The methodscould be improved by adding additional field recordsof site scale variables.

Physical habitat evaluation applying the AQEMsite protocol provides the specific information for nu-merous hydromorphological variables, such as thenumber of organic substrates, the amount of woodydebris (debris dams, logs), maximum current velocit-ies and the coefficient of variation (CV) of channeldepth, that directly or indirectly influence the in-stream biocoenosis. These mesohabitat characteristicshave previously been reported as important descriptorsof the macroinvertebrate community structure (Beiselet al., 1998).

Development of a Structure Index for mid-sized sandbottom streams in the German lowlands

The results presented in this study stress the im-portance of environmental variables for the develop-ment and implication of tools to assess river healthin Europe. However, future assessment systems forEuropean streams and rivers should predominantly bebased on riverine biota (EU commission, 2000). TheWFD has designated several Biological Quality Ele-ments (BQE; e.g., fish, benthic macroinvertebrates)instead of abiotic factors, such as physical habitatcharacteristics, to be predominantly used for assess-ment. The results of this study on the potential ofhydromorphological variables to detect and describehydromorphological degradation, therefore, have tobe integrated with a system that is based on bio-coenotic measures of the riverine community. Thiswas achieved by combining eight groups of hydro-morphological variables (woody debris, organic sub-strates, shading, shoreline, positive and negative struc-ture elements, land use, scouring, and bank fixation)to a newly developed measure, the German Struc-ture Index (GSI). Finally, single community measures(metrics, e.g., feeding types, current preferences, sub-strate preferences) and single indicator taxa can beidentified to provide candidate metrics of a multi-metric index to assess the ecological quality of a site(Hering et al., 2004). Lorenz et al. (2004) documentedthe interdependence between the hydromorphologicalquality of a site and numerous metrics derived fromthe macroinvertebrate community sampled at that site.Feld et al. (2002) found, for example, the number ofSimuliid taxa to be significantly higher at hydromor-phologically ‘unstressed’ sites.

In comparison to the existing methods of phys-ical habitat evaluation (e.g., the German ‘Struktur-gütekartierung’; LAWA, 2000), the GSI provides twoadvantages: First, on a numerical scale, the GSI isa continuous measure of hydromorphological qual-ity, allowing of simple correlation with biocoenoticmetrics. Second, the development of the GSI refersto hydromorphological reference conditions, whichrepresent one end of the hydromorphological gradi-ent. Thus, even if the pre-classified reference sitesare already influenced by slight hydromorphologicaldegradation, the variables identified to describe the re-spective end of the gradient are likely to be the samevariables suited to describe the reference conditions.

A potential deficit of the AQEM approach was thesubjective pre-selection of candidate sites according tothe researcher’s subjective judgment on the stressor-specific ecological status of a site. This approach waschosen to cover the whole gradient of the presumedmain stressor as good as possible. This is arguablya prerequisite for the detection of a gradual impactof this stressor. For German stream types, the mainstressor appears to be hydromorphological degrada-tion; organic pollution and acidification play a minorrole (e.g., HMULF, 1999; MUNLV/LUA NRW, 2000;NLÖ, 2000). Acidification can be objectively meas-ured, however, hydromorphological quality is ratherdifficult to scale. The GSI represents a method tomeasure hydromorphological degradation based onobjectively recorded hydromorphological attributes.Numerous site protocol variables clearly describedhigh quality and poor or bad quality sites (Fig. 5).They revealed an obvious gradient, which even al-lowed the establishment of a five-class classificationsystem (Fig. 2). Some subjectivity remained in thedefinition of quality groups necessary for ‘IndVal’analysis. However, this was ‘objectified’ by using onlythe extremes of the hydromorphological gradient re-vealed by NMS ordination (i.e., reference and heavilydegraded).

Lorenz et al. (2004), Feld et al. (2002) and Paulset al. (2002) reported the GSI a suitable measure forthe identification of biocoenotic metrics to assess theimpact of hydromorphological degradation on benthicmacroinvertebrates.

Acknowledgements

I would like to thank Dr V. W. Framenau, WesternAustralian Museum, Perth, Australia, for numerous

86

valuable comments and linguistic revision of the ma-nuscript. Dr Daniel Hering, University of Essen, Ger-many, provided valuable comments that contributedto this paper. Many thanks to Hanneke Vlek, Alterra,Wageningen, The Netherlands, and an anonymous re-viewer, who helped to improve the manuscript bynumerous important remarks and critical comments.I am also grateful to Melissa L. Thomas, University ofCalifornia, San Diego, U.S.A., for valuable commentson the manuscript.

AQEM was funded by the European Commission,5th Framework Program, Energy, Environment andSustainable Development, Key Action Water, Contractno. EVK1-CT1999-00027.

References

Addinsoft SARL, 2002. XLSTAT 5.2. Addinsoft, Paris.Agence-de-l’Eau-Rhin-Meuse, 1996. Outil d’évaluation de la qual-

ité du milieu physique. Metz.Allan, J. D., D. L. Erickson & J. Fay, 1997. The influence on catch-

ment land use on stream integrity across multiple spatial scales.Freshwat. Biol. 37: 149–161.

AQEM consortium, 2002. Manual for the application of the AQEMsystem. A comprehensive method to assess European streamsusing benthic macroinvertebrates, developed for the purpose ofthe Water Framework Directive. EVK1-CT1999-00027, Version1.0. Available via the Internet from www.aqem.de.

Barbour, M. T., J. Gerritsen, B. D. Snyder & J. B. Stribling, 1999.Rapid bioassessment protocols for use in streams and wade-able rivers: periphyton, benthic macroinvertebrates and fish. EPA841-B-99-002, U.S. Environmental Protection Agency, Office ofWater, Washington D.C.

Beisel, J.-N., P. Usseglio-Polatera, S. Thomas & J.-C. Moreteau,1998a. Effects of mesohabitat sampling strategy on the assess-ment of stream quality with benthic invertebrate assemblages.Arch. Hydrobiol. 142: 493–510.

Beisel, J.-N., P. Usseglio-Polatera, S. Thomas & J.-C. Moreteau,1998b. Stream community structure in relation to spatial vari-ation: the influence of mesohabitat characteristics. Hydrobiolo-gia 389: 73–88.

Benke, A. C., R. L. Henry III., D. M. Gillespie & R. J. Hunter, 1985.Importance of snag habitat for animal production in southeasternstreams. Fisheries 10: 8–13.

Benke, A. C., G. E. Willeke, F. K. Parrish & D. L. Stites, 1981. Ef-fects of urbanization on stream ecosystems. A-005-GA, Office ofWater Research and Technology, U.S. Department of the Interior.

Bundesanstalt für Geowissenschaften und Rohstoffe (FederalAgency of Geosciences and Raw Materials, Germany), 1993.Geologische Karte der Bundesrepublik Deutschland. Bundesan-stalt für Geowissenschaften und Rohstoffe, Hannover.

Clarke, K. R., 1993. Non-parametric multivariate analysis ofchanges in community structure. Aust. J. Ecol. 18: 117–143.

CSN 75 7221, 1998. Water Quality – Classification of Surface Wa-ter Quality. Czech Technical State Standard. Czech StandardsInstitute, Prague, 10 pp.

Dahl, J., R. K. Johnson & L. Sandin, 2004. Detection of or-ganic pollution of streams in southern Sweden using benthicmacroinvertebrates. Hydrobiologia 516: 161–172.

Davies, N. M., R. H. Norris & M. C. Thomas, 2000. Predictionand assessment of local stream habitat features using large-scalecatchment characteristics. Freshwat. Biol. 45: 343–369.

DEV (Deutsches Institut für Normung e. V.) (German Institute ofStandardization), 1992. Biologisch-ökologische Gewässerunter-suchung: Bestimmung des Saprobienindex (M2). In: DeutscheEinheitsverfahren zur Wasser-, Abwasser- und Schlammunter-suchung. VCH Verlagsgesellschaft, Weinheim, 13 pp.

Dudley, T. & N. H. Anderson, 1982. A survey of invertebrates as-sociated with wood debris in aquatic habitats. Melanderia 39:1–21.

Dufrêne, M. & P. Legendre, 1997. Species assemblages and indic-ator species: the need for a flexible asymmetrical approach. Ecol.Monogr. 67(3): 345–366.

Ellenberg, H., 1996. Vegetation Mitteleuropas mit den Alpen in öko-logischer, dynamischer und historischer Sicht. Ulmer, Stuttgart,1095 pp.

EU commission, 2000. Directive 2000/60/EC of the European Par-liament and of the Council of 23 October 2000 establishing aframework for Community action in the field of water policy.Official Journal of the European Communities, Brussels, 72 pp.

Feld, C. K., E. Kiel & M. Lautenschläger, 2002. The indica-tion of morphological degradation of streams and rivers usingSimuliidae. Limnologica 32: 273–288.

Franquet, E., S. Dolédec & D. Chessel, 1995. Using multivari-ate analyses for separating spatial and temporal effects withinspecies-environment relationship. Hydrobiologia 300/301: 425–431.

Frissell, C. A., W. J. Liss, C. E. Warren & M. D. Hurley, 1986. Ahierarchical framework for stream habitat classification: viewingstreams in a watershed context. Environ. Manage. 10: 199–214.

Gurnell, A. M., K. J. Gregory & G. E. Petts, 1995. The role ofcoarse woody debris in forest aquatic habitats: implications formanagement. Aquat. Cons. Mar. Freshwat. Ecosyst.: 143–166.

Harmon, M. E., J. F. Franklin, F. J. Swanson, P. Sollins, S. V.Gregory, J. D. Lattin, N. H. Anderson, S. P. Cline, N. G. Auman,J. R. Sedell, G. W. Lienkaemper, K. C. Jr. & K. W. Cummins,1986. Ecology of coarse woody debris in temperate ecosystems.Adv. ecol. Res. 15: 133–302.

Hering, D. & M. Reich, 1997. Bedeutung von Totholz fürMorphologie, Besiedlung und Renaturierung mitteleuropäischerFließgewässer. Nat. Landsch. 72: 383–389.

Hering, D., O. Moog, L. Sandin & P. F. M. Verdonschot, 2004.Overview and application of the AQEM assessment system.Hydrobiologia 516: 1–20.

Hering, D., A. Buffagni, O. Moog, L. Sandin, M. Sommerhaeuser, I.Stubauer, C. K. Feld, R. Johnson, P. Pinto, N. Skoulikidis, P. Ver-donschot & S. Zahrádková, 2003. The development of a systemto assess the ecological quality of streams based on macroinver-tebrates – design of the sampling programme within the AQEMproject. Int. Rev. Hydrobiol. 88: 345–361.

Hildrew, A. G., 1996. Whole river ecology: spatial scale heterogen-eity in the ecology of running waters. Arch. Hydrobiol., Suppl.113 (Large Rivers 10): 25–43.

HMULF (Regional Ministry of Environment, Agriculture andForestry of Hassia, Germany) (ed.), 1999. Gewässerstrukturgütein Hessen 1999. Wiesbaden, 52 pp.

Hoffmann, A. & D. Hering, 2000. Wood-associated macroinverteb-rate fauna in Central European streams. Int. Rev. Hydrobiol. 85:25–48.

Hughes, R. M., 1995. Defining acceptable biological status by com-paring with reference conditions. In: Davis, W. S. & T. P. Simon(eds), Biological Assessment and Criteria. Tools for Water Re-

87

source Planning and Decision Making. Lewis Publishers, BocaRaton, 415 pp.

Illies, J., 1978. Limnofauna Europaea. Gustav Fischer, Stuttgart,532 pp.

Jones, R. C. & C. C. Clark, 1987. Impact of watershed urbanizationon stream insect communities. Water Resour. Bull. 23: 1047–1055.

LAWA (Federal State’s Working Group ‘Water’, Germany)(ed.), 2000. Gewässerstrukturgütekartierung in der Bundes-republik Deutschland – Verfahren für kleine und mittelgroßeFließgewässer. Kulturbuchverlag, Schwerin, 145 pp + appen-dices.

Lorenz, A., D. Hering, C. K. Feld & P. Rolauffs, 2004. A newmethod for assessing the impact of hydromorphological degrada-tion on the macroinvertebrate fauna of five German stream types.Hydrobiologia 516: 107–127.

LUA NRW (Environmental Agency of North Rhine-Westphalia,Germany) (ed), 2001. Leitbilder für die mittelgroßen bisgroßen Fließgewässer in Nordrhein-Westfalen – Flusstypen.LUA Merkbl. 34, Essen, 130 pp.

LUA NRW (Environmental Agency of North Rhine-Westphalia,Germany) (ed), 2002. Fließgewässertypenatlas Nordrhein-Westfalens, LUA Merkbl. 36, Essen, 61 pp.

McCune, B. & M. J. Mefford, 1999. Multivariate Analysis of Eco-logical Data (PCOrd), Version 4.25. MjM Software, GlenedenBeach, Oregon, U.S.A.

MUNLV/LUA NRW (Regional Ministry of Environment andNature Protection, Agriculture and Consumer Affairs NorthRhine-Westphalia & Regional Environmental Agency of NorthRhine-Westphalia, Germany) (eds), 2000. Gewässergüte-bericht 2000 – 30 Jahre biologische Gewässerüberwachung inNordrhein-Westfalen. Düsseldorf, 346 pp.

Mutz, M., 2000. Influence of woody debris on flow patterns andchannel morphology in a low energy, sand-bed stream reach. Int.Rev. Hydrobiol. 85: 107–121.

Newbold, J. D., D. C. Erman & K. B. Roby, 1980. Effects of loggingon macroinvertebrates in streams with and without buffer strips.Can. J. Fish. aquat. Sci. 37: 1076–1085.

Nijboer, R. C., R. K. Johnson, P. F. M. Verdonschot, M. Sommer-häuser & A. Buffagni, 2004. Establishing reference conditionsfor European streams. Hydrobiologia 516: 91–105.

NLÖ (Regional Agency of Ecology Lower Saxony, Germany) (ed.),2000. Gewässergütebericht 2000. Hildesheim, 40 pp.

Ofenböck, T., O. Moog, J. Gerritsen & M. Barbour, 2004. A stressorspecific multimetric approach for monitoring running waters inAustria using benthic macro-invertebrates. Hydrobiologia 516:251–268.

Omernik, J. M., 1987. Ecoregions of the conterminous United States(with map). Ann. Assoc. am. Geogr. 77: 118–125.

Pauls, S., C. K. Feld, M. Sommerhäuser & D. Hering, 2002. NeueKonzepte zur Bewertung von Tieflandbächen und -flüssen nach

Vorgaben der EU Wasser-Rahmenrichtlinie. Wasser Boden 54:70–77.

Podani, J., 2000. Introduction to the Exploration of MultivariateBiological Data. Backhuys Publishers, Leiden, 407 pp.

Rabeni, C. F., 2000. Evaluating physical habitat integrity in relationto the biological potential of streams. Hydrobiologia 422/423:245–256.

Raven, P. J., N. T. H. Holmes, P. Charrier, F. H. Dawson, M.Naura & P. J. Boon, 2002. Towards a harmonized approach forhydromorphological assessment of rivers in Europe: a qualit-ative comparison of three survey methods. Aquat. Cons. Mar.Freshwat. Ecosyst. 12: 405–424.

Raven, P. J., N. T. H. Holmes, F. H. Dawson & M. Everard, 1998.Quality assessment using river habitat survey data. Aquat. Cons.Mar. Freshwat. Ecosyst. 8: 477–499.

Raven, P. J., P. Fox, M. Everard, N. T. H. Holmes & F. H. Dawson,1997. River Habitat Survey: a new system for classifying riversaccording to their habitat quality. In Boon, P. J. & D. L. How-ell (eds), Freshwater Quality: Defining the Indefinable. ScottishNatural Heritage, Edinburgh: 215–234.

Richards, C., L. B. Johnson & G. E. Host, 1996. Landscape-scaleinfluences on stream habitats and biota. Can. J. Fish. aquat. Sci.53 (Suppl. 1): 295–311.

Rolauffs, P., I. Stubauer, Z. Zahrádková, K. Brabec & O. Moog,2004. Integration of the saprobic system into the European UnionWater Framework Directive. Hydrobiologia 516: 285–298.

Sponseller, R. A., E. F. Benfield & H. M. Valett, 2001. Relationshipsbetween land use, spatial scale and stream macroinvertebratecommunities. Freshwat. Biol. 46: 1409–1424.

Statistisches Bundesamt (Federal Agency of Statistics, Germany)(ed.), 1997. Daten zur Bodenbedeckung. Wiesbaden.

Statzner, B., B. Bis, S. Doledec & P. Usseglio-Polatera, 2001.Perspectives for biomonitoring at large spatial scales: a unifiedmeasure for the functional composition of invertebrate com-munities in European running waters. Basic Appl. Ecol. 2:73–85.

Tabacchi, E., D. L. Correll, R. Hauer, G. Pinay, A.-M. Planty-Tabacchi & R. C. Wissmar, 1998. Development, maintenanceand role of riparian vegetation in the river landscape. Freshwat.Biol. 40: 497–516.

Vlek, H. E., P. F. M. Verdonschot & R. C. Nijboer, 2004. Towardsa multimetric index for the assessment of Dutch streams usingbenthic macroinvertebrates. Hydrobiologia 516: 173–189.

Whittier, T. R., R. M. Hughes & D. P. Larsen, 1988. Correspondencebetween ecoregions and spatial patterns in stream ecosystems inOregon. Can. J. Fish. aquat. Sci. 45: 1264–1278.

Wiederholm, T. & R. K. Johnson, 1997. Monitoring and assess-ment of lakes and watercourses in Sweden. In Ottens, J. J.,F. A. M. Claessen, P. G. Stoks, J. G. Timmerman & R. C.Ward (eds), Monitoring Tailor-made II, Information Strategiesin Water, Nunspeet, The Netherlands: 317–329.

88

Appendix 1. List of site protocol variables with notes on numerical and spatial scale. Variable usage for different multivariateanalysis is indicated by a ‘+’, exclusion from analysis by a ‘−’. Numerical scale assigned according to Podani (2000). Areal andlongitudinal extent of spatial scale is explained in Chapter ‘Evaluation of hydromorphological degradation’.

Variable Variable name Numerical Spatial Typo- Degradation

code scale scale logy All German Stream

stream stream type

types types D03

7 Stream order (Strahler system) Ordinal Catchment + + + +8 Distance to source (km) Interval Catchment + + + +11 Altitude (m a.s.l.) Interval Catchment + + + +12 Ecoregion (according to Illies, 1978) Nominal Catchment + + + −15 Catchment area (km2) Interval Catchment + + + +16 Size typology according to the WFD (EU commission, 2000) Ordinal Catchment + + + −17 Stream density (km km−2) Interval Catchment + + + +18–1 Geology: Acid silicate rocks (%) Ratio Catchment + + + −18–3 Geology: Carbonate rocks (%) Ratio Catchment + + + +18–4 Geology: Alluvial deposits (%) Ratio Catchment + + + −18–7 Geology: Moraines (%) Ratio Catchment + + + +18–8 Geology: Sander (%) Ratio Catchment + + + +18–9 Geology: Marine deposits (%) Ratio Catchment + + + −18–10 Geology: Organic formations (%) Ratio Catchment + + + +18–11 Geology: Loess (%) Ratio Catchment + + + +18a Geological typology (silicate, carbonate, organic) Ratio Catchment + + + +19–91 Land use: Native forest (%) Ratio Catchment + + + +19–4 Land use: Wetland (mire) (%) Ratio Catchment + + − −19–5 Land use: Open grass-/bush land (%) Ratio Catchment + + + +19–9 Land use: Artificial standing water bodies (ponds, etc.) (%) Ratio Catchment + + + +19–10 Land use: Non-native forest (%) Ratio Catchment + + + +19–12 Land use: Crop land (%) Ratio Catchment + + + +19–13 Land use: Pasture (%) Ratio Catchment + + + +19–92 Land use: Total agriculture (%) Ratio Catchment + + + −19–15 Land use: Urban sites (residential) (%) Ratio Catchment + + + +24 Hydrologic stream type (permanent, Nominal Catchment + + + −

periodic/intermittent, episodic)

25 Presence of lakes in the whole upstream Binary Catchment + + + +continuum

26 Width of the floodplain (m) Interval Site − − − +29 Valley shape (V-shaped, U-shaped, Nominal Site + + + −

trough, meander valley, etc.)

30–91 Land use: Native forest (%) Ratio Site − − − +30–92 Land use: Open grass-/bush land, reeds (%) Ratio Site − − − +30–10 Land use: Non-native forest (%) Ratio Site − − − +30–12 Land use: Crop land (%) Ratio Site − − − +30–13 Land use: Pasture (%) Ratio Site − − − +30–93 Land use: Total agriculture (%) Ratio Site − − − +30–15 Land use: Urban sites (residential) (%) Ratio Site − − − +31 Number of other transverse structures Interval Upstream + + + +34 Straightening Binary Upstream + + + +

Continued on p. 89

89

Appendix 1. Continued.

Variable Variable name Numerical Spatial Typo- Degradation

code scale scale logy All German Stream

stream stream type

types types D03

35 Removal of coarse woody debris (CWD) Binary Upstream + + + +36 Cut-off meanders Binary Upstream + + + +37 Scouring below bank top (m) Interval Upstream + + + +38 Culverting Binary Upstream + + + +39 Number of other transverse structures Interval Downstream + + + +42 Straightening Binary Downstream + + + +43 Removal of coarse woody debris (CWD) Binary Downstream + + + +44 Cut-off meanders Binary Downstream + + + +45 Scouring below bank top (m) Interval Downstream + + + +46 Culverting Binary Downstream + + + +47 Number of dams retaining sediment Interval Upstream + + + +49 Number of dams obstructing migration Interval Downstream + + + +56 Impoundments or dams (% of length) Ratio Upstream + + + +56a Lack of natural wooded vegetation Binary Upstream + + + +56b Non-native wooded vegetation Binary Upstream − − − +57 Lack of natural wooded vegetation Binary Downstream + + + +58 Non-native wooded vegetation Binary Downstream − − − +59 Impoundments or dams (% of length) Ratio Downstream + + + +61 Non-source pollution Binary Upstream + + + −63 Eutrophication Binary Upstream + + + −68 Mean depth at bankfull discharge (m) Interval Site + + + +69 Shading at zenith (foliage cover) (%) Ratio Site + + + +70-91 Average width of wooded riparian Interval Site + + + +

vegetation right + left (m)

71 Channel form (braided, meandering, sinuate, etc.) Nominal Site + + + +73 Presence of natural standing water bodies Binary Site + + + +

in the floodplain (e.g. backwaters)

74 Number of debris dams > 0.3 m3 Interval Site + + + +75 Number of logs > 10 cm diameter Interval Site + + + +76–91 Shoreline covered with wooded riparian vegetation right + left (%) Ratio Site + + + +77 Number of dams Interval Site + + + +78 Number of other transverse structures Interval Site + + + +79–91 Bank fixation stones (rip rap) (%) Ratio Site + + + +79–92 Bank fixation wood/trees (%) Ratio Site + + + +79–93 No bank fixation (%) Ratio Site + + + +80–3 Bed fixation stones (%) Ratio Site + + + +80–9 No bed fixation (%) Ratio Site + + + +81 Stagnation Binary Site + + + +84 Straightening Binary Site + + + +85 Removal of coarse woody debris (CWD) Binary Site + + + +86 Cut-off meanders Binary Site + + + +87 Scouring below bank top (m) Interval Site + + + +88 Culverting Binary Site + + + +92 Impoundments at sampling site Binary Site + + + +

Continued on p. 90

90

Appendix 1. Continued.

Variable Variable name Numerical Spatial Typo- Degradation

code scale scale logy All German Stream

stream stream type

types types D03

93 Removal/lack of natural floodplain vegetation Binary Site + + + +94 Non-native wooded riparian vegetation Binary Site − − − +95 Source pollution Binary Site + + + −96 Non-source pollution Binary Site + + + −97 Sewage overflows Binary Site + + + −98 Eutrophication Binary Site + + + −103–2 Megalithal (> 40 cm) (%) Ratio Site + + − −103–3 Macrolithal (> 20 cm to 40 cm) (%) Ratio Site + + + +103–4 Mesolithal (> 6 cm to 20 cm) (%) Ratio Site + + + +103–5 Microlithal (> 2 cm to 6 cm) (%) Ratio Site + + + +103–6 Akal (> 0.2 cm to 2 cm) (%) Ratio Site + + + +103–7 Psammal/psammopelal (%) Ratio Site + + + +103–8 Argyllal (< 6 µm) (%) Ratio Site + + + −104–2 Algae (%) Ratio Site + + + +104–3 Submerged macrophytes (%) Ratio Site + + + +104–4 Emergent macrophytes (%) Ratio Site + + + +104–5 Living parts of terrestrial plants (%) Ratio Site + + + +104–6 Xylal (wood) (%) Ratio Site + + + +104–7 CPOM (%) Ratio Site + + + +104–8 FPOM (%) Ratio Site + + + +104–10 Organic mud, sludge (%) Ratio Site + + + +104–11 Debris (e.g. empty mollusc shells at the shore zone) (%) Ratio Site − − − +104–91 Number of organic substrates Interval Site + + + +105 Average stream width (m) Interval Site + + + +110 pH value Interval Site + + + −111 Conductivity (µS cm−1) Interval Site + + + −112 Reduction phenomena Binary Site + + + −113 Waste Binary Site + + + −114 Dissolved oxygen content (mg l−1) Interval Site + + + −118 Max. depth (cm) Interval Site − − − +120 Max. current velocity (m s−1) Interval Site − − − +121 Mean depth (cm) Interval Site + + + +122 CV depth Ratio Site + + + +123 Mean current velocity (m s−1) Interval Site + + + +124 CV current velocity Ratio Site + + + +125 Ammonium (mg l−1) Interval Site + + + −127 Nitrate (mg l−1) Interval Site + + + −128 Ortho phosphate (µg l−1) Interval Site + + + −129 Total phosphate (µg l−1) Interval Site + + + −