ARTICLE IN PRESS

Limnologica 36 (2006) 119–137

www.elsevier.de/limno

The GW-Fauna-Index: A first approach to a quantitative ecological

assessment of groundwater habitats
Hans Jürgen Hahn
Universität Koblenz-Landau, Campus Landau, Abteilung Biologie, Im Fort 7, D-76829 Landau, Germany

Received 13 April 2005; received in revised form 9 January 2006; accepted 8 February 2006

Abstract

Between June 2001 and December 2002, 18 hyporheic and groundwater bores were sampled for fauna and
environmental data using phreatic traps. The bores were situated in three different natural geographic regions in
Palatinate, Southwestern Germany.
Faunal data correlated with the relative amount of detritus, bacterial abundances and the standard deviation of
temperature, while very few and weak correlations were found with physical–chemical variables. Dissolved oxygen was
assumed to be a limiting factor for most metazoans with a critical concentration at around 0.5–1 mg l
1.
To quantify the strength of the hydrological exchange with surface water and its effects on fauna, a so-called GW-
Fauna-Index was developed and calculated using the relative amount of detritus, standard deviation of temperature,
and oxygen concentration. From all environmental data and on all spatial scales, this index best explained the total
faunal abundance and taxonomic richness.
To describe the availability of organic aliments in the groundwater, the terms of ‘‘alimonic’’ and ‘‘alimony’’ [from
lat. alimonium ¼ (food) supply] were proposed.
Although stygofauna was different in the geographic regions investigated, the GW-Fauna-Index was independent
from these regional particularities. Using the GW-Fauna-Index, three groups of groundwater habitats could be
classified according to the alimonic conditions. From oligo-alimonic group I samples, fauna was mostly absent, while
meso-alimonic group II samples were prevailingly populated by stygobites, and eu-alimonic group III samples by
ubiquists and stygoxenes. Total abundances and taxonomic richness increased significantly from group I to group III.
Group I samples were characterized by low index values, group II samples by intermediate and group III samples by
high values.
The GW-Fauna-Index provides promising perspectives for application, but needs some improvement. First of all,
detritus should be analysed quantitatively and qualitatively, rather than semi-quantitatively. Also, a standard protocol
for sampling has to be developed.
r 2006 Elsevier GmbH. All rights reserved.

Keywords: Stygofauna; Ecological assessment; Groundwater habitats; GW-Fauna-Index; Trophic classification; Alimony

Introduction
Tel.: +49 6341 280 211; fax: +49 6341 280 341. All groundwater biologists have the problem of ‘‘bad
E-mail address: hjhahn@uni-landau.de. predictability’’ to contend with: very often, bores, which

0075-9511/$ - see front matter r 2006 Elsevier GmbH. All rights reserved.
doi:10.1016/j.limno.2006.02.001
 ARTICLE IN PRESS
120 H.J. Hahn / Limnologica 36 (2006) 119–137

are situated close together and which seem to be very fauna (e.g. Dumas, Bou, & Gibert, 2001; Hahn &
similar in construction, depth and hydrogeology, Friedrich, 1999; Plénet, Gibert, & Marmonier, 1995). To
harbour a different fauna. As an extreme example, a a certain degree, oxygen is an exception, but the
bore could be well colonized on one sampling occasion, available information demonstrating its importance is
but fauna may be completely absent another time. In my contradictory. Malard and Hervant (1999) compared
experience, around 30% of all groundwater samples in several studies, of which some found a strong correla-
any area and on different scales are free of metazoans tion between fauna and oxygen, while others detected
(see also Fuchs & Hahn, 2004). Evidently, distribution none. Strayer et al. (1997) observed good correlations in
of stygofauna is very patchy (Mösslacher, 2003) and the hyporheic zone between fauna and organic matter
dynamic through space and time. In other words: even only, if oxygen concentration were41 mg l
1. Haken-
over short periods of time and on a local scale of some kamp and Palmer (2000) argue that oxygen is important
meters, the variation of species and communities can be for fauna, particularly when in lower concentrations.
strong – an indication of the heterogenity and the Evidently, oxygen is a limiting factor in the ground-
dynamic nature of groundwater habitats as mentioned water.
by numerous authors (e.g., Gibert, 2001; Gibert & Since groundwater is mostly a heterotrophic habitat,
Deharveng, 2002; Griebler & Mösslacher, 2003; Hahn, food and oxygen availability is determined by import
2004; Hahn & Matzke, 2005; Mösslacher, 1998; from the surface. This so-called hydrological exchange
Mösslacher, 2003). seems to be the key factor shaping most groundwater
The occurrence of stygofauna depends on food supply communities (e.g. Dôle-Olivier & Marmonier, 1992;
in the form of detritus and bacteria, but also on the Dumas et al., 2001; Griebler & Mösslacher, 2003; Hahn
competition with stygoxenous species. The effects of & Friedrich, 1999; Hakenkamp & Palmer, 2000;
stygoxenes on groundwater communities increase with Malard, Ferreira, Dolédec, & Ward, 2003). Together
increasing food supply, for example as a result of with oxygen and organic matter (DOM and POM)
organic pollution (Malard, Mathieu, Reygrobellet, & stygoxenous fauna also have access to the groundwater.
Lafont, 1996, 1999; Malard, Reygrobellet, Mathieu, & If food and oxygen supply are sufficently high, these
Lafont, 1994; Sket, 1999). There are some indications species are able to durably colonize the groundwater
for the importance of organic matter (Bretschko & and compete out the stygobites (Malard et al., 1994,
Leichtfried, 1988; Brunke & Gonser, 1999; Datry, 1996, 1999; Sket, 1999).
Malard, & Gibert, 2005; Gibert & Deharveng, 2002: Strength and direction of the hydrological exchange
Griebler & Mösslacher, 2003; Hahn & Matzke, 2005; depend on both local (soils, geomorphology, sediment
Strayer, May, Nielsen Wolheim, & Hausam, 1997) and structure, hydraulic and depth of groundwater table)
bacteria (Brown, Rundle, Hutchinson, Williams, & and landscape (hydrogeology, climate) factors.
Jones, 2003; Brunke & Fischer, 1999; Griebler & Roughly, we can estimate the strength of the hydro-
Mösslacher, 2003; Strayer, 1994) both in the ground- logical exchange by using the amplitude of water
water and in the hyporheic zone, but on the whole, the temperature: while the temperature of groundwater is
available data on food interactions in the groundwater is nearly constant and corresponds to the annual mean air
scant (Rumm, 1999). In the hyporheic zone, species temperature (Dreher, Pospisil, & Danielopol, 1997), the
richness and the abundance of invertebrate fauna are temperature of surface waters varies according to the
often positively correlated with bacterial abundance actual air temperature. In general, one can say the
(Brunke & Fischer, 1999) and organic matter (Strayer et higher the amplitude of temperature in the groundwater
al., 1997). Similar results were found for the ground- the stronger the surface water influence as a result of the
water, where fauna abundance and taxonomic richness hydrological exchange (Schönborn, 2003).
were correlated with dissolved organic matter (DOM) The problem of ‘‘bad predictability’’ is a severe
(Datry et al., 2005). However, Storey and Williams hindrance to the investigation and evaluation of
(2004) found no correlations, with neither bacterial groundwater habitats. This problem is likely to be
abundance and activity nor with particulate organic linked to the food and oxygen supply and thus to the
matter (POM). The authors suggest that it was a result hydrological exchange. If it would be possible to
of low oxygen concentrations. describe the hydrological exchange quantitatively, it
Sediment structure determines the hydraulic conduc- also should be possible to make some predictions on the
tivity and the size of the available living space, with character of the groundwater fauna, and this would be a
higher abundances in coarse compared to fine sediments first step to a quantitative ecological assessment of
(Strayer et al., 1997, Mösslacher, 1998). Larger pore groundwater habitats.
spaces enhance groundwater flow and thus improve the In this paper, I examine how environmental condi-
connectivity between different groundwater habitats and tions in the groundwater affect the fauna communities
surface water. Hydrochemical parameters seem to have based on an estimation of the hydrological exchange. To
no, or only weak effects on groundwater and hyporheic achieve this, a preliminary index (the GW-Fauna-Index)
 ARTICLE IN PRESS
H.J. Hahn / Limnologica 36 (2006) 119–137 121

is calculated from the oxygen concentration, the relative aquifers. Typical of the alluvial groundwater is that it is
amount of detritus, and the standard deviation of recharged by the adjacent fractured rock aquifers, but
temperature. These factors were used, because oxygen not directly from soil surface (Hahn, 1996). ‘‘Haard-
concentration and detritus are considered as limiting trand’’ is the name of the western fault belt of the Upper
factors and dependent upon the hydrological exchange, Rhine Valley with an extremely heterogenous geology.
which is approximately reflected by the standard The typical landuse is an intensive viticulture. Miner-
deviation of temperature. alization of the Haardtrand groundwater is higher than
in the Pfälzerwald Mountains. Due to the fractured
aquifers, groundwater tables at the Haardtrand study site
Study area varied greatly (several meters), within the study period.
In the Upper Rhine Plateau, most of the soils are fertile
The study area was situated in Palatinate, South- and intensively used for agriculture, but the poor, sandy
western Germany, and comprised three different natural soils of the study site of Offenbacher Wald are
geographic regions (‘‘Naturraum’’, see Hahn, 2002): completely forested. The groundwater in those floodplain
Pfälzerwald Mountains, Haardtrand and Upper Rhine forests is well mineralized. Annual mean air temperature
Plateau (Fig. 1). Three hyporheic and 15 alluvial is relatively low in the Pfälzerwald mountains (8.6 1C),
groundwater bores (each of the groundwater bores with but increases eastward with 9.7 1C in the Haardtrand and
three traps, see method section) were sampled monthly 10.1 1C in the Upper Rhine Plateau (Geiger, Preuss, &
from June 2001 to December 2002. The bores were Rothenberger, 1991; Geiger, 1996).
situated in four alluvial floodplains, with the Kolbental
and Klammtal Valleys in the Pfälzerwald Mountains,
the Modenbachtal Valley in the Haardtrand, and Methods
Offenbacher Wald on the Upper Rhine Plateau.
The Pfälzerwald Mountains are a forested sandstone Fauna, water and bacteria samples were pumped
area with poor, sandy soils and a high rate of ground- from traps, which are described in detail by Hahn
water recharge. Its groundwater is characterized by a low (2005). The depth of the groundwater bores was 5.0 m (3
degree of mineralization and high oxygen concentrations. bores) and 7.50 m (12 bores). The three traps of each
It flows eastward along the alluvials of local brooks via bore were installed 1.70 m (trap A, near the groundwater
the Haardtrand and recharges the Upper Rhine Plateau table) and 2.90 m (trap B) below soil surface, and the

Fig. 1. Map of the study area. indicates the sample sites. All sites have 4–5 transsectional bores.
 ARTICLE IN PRESS
122 H.J. Hahn / Limnologica 36 (2006) 119–137

third trap C, 50 cm above the bottom of the bores Dissolved oxygen (DO), electrical conductance (EC),
(Fig. 2). The three hyporheic bores contained one trap, temperature and pH-values were measured using a
which was fixed 30 cm below sediment surface. Prior to WTW multimeter Multiline P4 (Wissenschaftlich-Tech-
pumping, the depth of the water table was measured. To nische Werkstätten GmbH, Weilheim) directly after
sample fauna, bacteria and sediments, the content of the having removed the water. Nitrate (NO
3 ), ammonium
traps (0.9 l of water) was pumped. Bacterial samples (NH+ 3

4 ), total dissolved iron and phosphate (PO4 ) of
were then separated from the main sample flow by a the sample were analysed using a Merck Reflectoquant
deviation between the pump and the traps. Thereafter, reflectometer, and buffering capacity using a Merck
another 2 l of water were sampled for the hydro- Aquamerck titration kit (Merck KGaA, Darmstadt).
chemical analyses. All samples were stored in a cold Chemical oxygen demand (COD), as a low priced
box. Animals were processed in the laboratory within substitute of dissolved organic carbon (DOC), was
24 h of collection. They were eluted, filtered with a measured using Dr. Lange test kits and a Dr. Lange
plankton net (63 mm) and sorted alive. Crustaceans were photometer (Dr. Lange GmbH, Düsseldorf; now Hach-
determined on species level and then classified either as Lange GmbH). Standard deviation of temperature (SD
stygobites or ubiquists/stygoxenes according to their temperature) was calculated for each trap using the field
ecological characterization by Einsle (1993), Janetzky, data over the study period.
Enderle, & Noodt (1996) and Meisch (2000). Stygophi- Bacterial samples were brought to a public labora-
lous species were generally classified in this study as tory, incubated at 20 and 36 1C according to the German
ubiquists/stygoxenes (see Table 3 in result section). Industrial Norm DIN 38 411, and colonies were counted
after an incubation period of 48 h. The colony numbers
measured using DIN 38 411 are not supposed to be
representative of groundwater bacteria, but to indicate
surface water affects on the groundwater. However, the
36 1C incubation counts correlated slightly better with
fauna data than the 20 1C incubation counts. Although
these differences were not significant on bore level (t-test
for paired samples: p40.1, n ¼ 18), in this study, the
results of the 36 1C incubation are presented. The
relative amount of detritus in the samples was estimated
on an ordinal scale (Table 1). Further analyses of the
detritus (e.g. determination of dry weight, C/N ratio,
microbial conditioning) could not be carried out for
methodical and financial reasons.
The GW-Fauna-Index was calculated using the
formula:
pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
GW
Fauna
Index ¼ DO ðmg=lÞ
pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
Relative Amount of Detritus  SD Temperature:

Table 1. Estimation of the relative amounts of sediment per

sample. From Hahn and Matzke (2005)

Scale Description Characterization

0 Absent No detritus in the sampling vessel

1 Little Bottom of the sampling vessel
(+ ¼ 7.6 cm) slightly covered by
detritus
2 Much Bottom of the sampling vessel
Fig. 2. The trap system used for this study. The traps were covered by several miilimetres of
fixed in the slotted area of the bores. Groundwater bores detritus
contained three traps, hyporheic bores one trap. Fauna and 3 Very much Bottom of the sampling vessel
sediments accumulate in the traps and are sampled by covered by one or more
pumping. Water and bacterial samples are pumped as well. centimetres of detritus
From Hahn (2005), modified.
 ARTICLE IN PRESS
H.J. Hahn / Limnologica 36 (2006) 119–137 123

These parameters were selected, because the relative

amount of detritus and the standard deviation of

amount of
PO4 (mg l
1) NH4 (mg l
1) Bacteria colony Relative

(colonies ml
1) detritus

temperature were best correlated with faunal data,

(0.61)

(1.07)

(0.76)
0.95

1.37

0.99
299

192
and because oxygen was considered to be a limiting

91
factor (see result section). SD temperature was con-
sidered as a direct indication of the hydrological
exchange and thus used as an unspecific corrective for

number

3000
lacking information on the nutritive quality of detritus

203

476
57

69

98
(see discussion section). Experiments with other for-
mulas, which also considered bacteria colony numbers
provided similar correlations (Spearman r) with total

(0.18)

(0.23)

(0.18)
0.32

0.33

0.33
238

116
abundance and numbers of taxa on bore level than the

61
formula displayed above (t-test for paired samples:
p40.5, n ¼ 17). Because of this, and because of the

(0.024)

(0.553)

(3.987)
difficulties associated with measuring bacterial abun-

0.016

0.192

3.129
269

129
63
dances in the groundwater (Griebler, 2003), bacteria
colony numbers were excluded from the formula. The

iron (mg l
1)

Total diss.
formula has been developed as an iterative process. The
three elements of the formula were weighted in different

(2.5)

(4.9)

(6.2)
264

151
1.0

2.7

7.6
71
ways. SD temperature was included without any
transformation, while square root transformation of

(mg l
1)

All values are indicated as mean values, with the exception of bacteria colony numbers, which are displayed as medians.
COD
DO and detritus yielded the best index correlations with

(4.1)

(7.8)

(7.9)
275

192
3.5

6.1

8.0
95
faunal data. Average values of environmental variables in the three geographical regions of the study area
Correlations were analysed by using the Spearman-
test. The Mann–Whitney U-test was used to identify
(mmol l
1)
Buffering
capacity
differences between unpaired groups. Differences be-

(0.3)

(0.4)

(1.0)
273

163
tween correlation coefficients of total abundance and
0.2

1.8

4.4
91
numbers of taxa with the GW-Fauna-Index and bacteria
(mS cm
1)

colony numbers, respectively, were analysed by the t-test

(169)
(32)

(51)
103

321

250

105

591

197
EC

for paired samples. Comparisons of the faunal commu-

nities between the traps were carried out by a multi-
pH-value

dimensional scaling (MDS) using the square root

(0.6)

(0.4)

(0.3)
300

206
transformed Bray-Curtis dissimilarity. To test the three
5.1

6.6

6.8
98

groups of groundwater habitats, a stepwise discriminant

(1C) at each trap
Temperature SD temperature

analysis was carried out. The predictor variables for

total abundance, numbers of taxa and for the two MDS
dimensions were identified by using a stepwise multiple
(0.56)

(0.69)

(0.97)
1.24

2.73

1.81

analysis of regression. Prior to statistical analyses, all

20

11

13

data were log (x+1) transformed. Data was processed

using SPSS 11.0 (SPSS Inc.), PRIMER 5.0 (Primer-E
Ltd.) and Excel 97 (Microsoft Corporation).
(1.7)

(3.1)

(2.3)
(1C)

10.8

11.2
309

206
9.2

97
NO3 (mg l
1)

Results
(3.4)

(3.4)
(4.6
282

162
5.7

3.6

2.8
90

Environmental variables
(mg l
1)

The groundwaters of the three geographical regions

(3.1)

(4.0)

(2.2)
DO

296

193
6.1

4.6

1.4
96

can be distinguished clearly by their physical–chemical

Pfälzerwald mountains

characteristics. In the Pfälzerwald Mountains DO was

Upper Rhine Plateau

high and nitrate moderate, while temperature, pH, EC,

buffering capacity, COD, phosphate and iron were low
Mean/median

Mean/median

Mean/median
Haardtrand
Table 2.

(Table 2). From the Pfälzerwald Mountains via the

Haardtrand to the Upper Rhine Plateau, DO and
(SD)

(SD)

(SD)

nitrate decreased, while temperature, pH, EC, buffering

n

n
 ARTICLE IN PRESS
124 H.J. Hahn / Limnologica 36 (2006) 119–137

capacity, COD, phosphate and iron increased. These destructor (Acari), respectively, occurred in the neigh-
differences were significant (U-test, po0:001) for all bouring bores HB1 and HB2. The Upper Rhine Plateau
parameters and between all geographical regions, with bores were characterized by a poor fauna count. With
the exception of temperature, which did not differ the exception of the hyporheic zone and the uppermost
significantly between the Haardtrand and the Upper trap of bore OW2, fauna was absent from many samples
Rhine Plateau (U-test: p ¼ 0:793). No differences at all or occurred in low numbers. The most common species
were found for NH+ 4 (U-test, p40.1). The standard were Diacyclops languidoides and Diacyclops bisetosus.
deviation of temperature at each trap was lowest in the Nematodes and oligochaetes occurred in all bores. All
Pfälzerwald Mountains and highest in the Haardtrand four Parastenocaris species reported from the area
with the values of the Upper Rhine Plateau falling (Hahn, 2004) were found. The Haardtrand bore fauna
between the two. These differences were significant (U- was very heterogenous, but in general, it was character-
test, po0:05). ized by high abundances of stygoxenes and ubiquists,
Similar to the SD of temperature, bacteria colony nematodes and oligochaetes. The taxonomic composi-
numbers were highest in the Haardtrand and lowest in tion is, according to the Haardtrand’s location between
the Pfälzerwald Mountains (Table 2) with significant Pfälzerwald Mountains and Upper Rhine Plateau, in
differences (U-test, po0:001) between all geographical between these geographical regions. For example,
regions. The relative amount of detritus was significantly stygobiontic Diacyclops languidoides c.f. and Niphargus
higher in the Haardtrand than in the Pfälzerwald aquilex, typical of the Pfälzerwald Mountains, were
Mountains (U-test: po0:002) and in the Upper Rhine recorded, and Parastenocaris fontinalis borea, the most
Plateau (U-test: p ¼ 0:007), while no differences were frequent and abundant parastenocarid in the Upper
found between Pfälzerwald Mountains and Upper Rhine Plateau, was the only parastenocarid found in the
Rhine Plateau (U-test: p ¼ 0:996). The parameters Haardtrand groundwater. Also, Graeteriella unisetigera,
which are related to food supply, were significantly a typical stygobite of the Palatine Upper Rhine Plateau,
correlated with the standard deviation of temperature of occurred in the Haardtrand.
each trap (n ¼ 44): bacteria colony number (r ¼ 0:816, On all spatial scales, several, mostly weak correlations
po0:0001), detritus (r ¼ 0:400, p ¼ 0:007) and COD were found between abiotic factors and fauna (total
(r ¼ 0:496, po0:001). Also, bacteria colony numbers abundance and numbers of taxa). Fauna had a strong
and detritus were weakly, but significantly correlated and positive correlation with the GW-Fauna-Index.
with each other (r ¼ 0:370, po0:001, n ¼ 328). No Also, significant positive correlations were detected
correlations were found between standard deviation of between fauna and detritus, bacteria colony numbers,
temperature and DO (r ¼
0:172, p ¼ 0:265, n ¼ 44) for and SD temperature. Total abundances increased with
the complete data set, but in the Pfälzerwald Mountains, increasing amounts of detritus and SD temperature.
there was a negative correlation (r ¼
0:605, p ¼ 0:005, (Fig. 3). The differences between the boxplot groups
n ¼ 20) and in the Upper Rhine Plateau a positive one were significant (U-test: p40.01). DO correlations with
(r ¼ 0:747, p ¼ 0:003, n ¼ 13). fauna, if existing, were incongruent, sometimes positive
and sometimes negative, depending on the geographical
region. For concentrations higher than 1.0 mg l
1, no
Fauna correlations (po0:05 and ro0:2) between DO and fauna
were found. Abundances per sample decreased strongly
General patterns below 1.0 mg l
1 (Fig. 3). The differences in metazoan
In the study area, 35 species and some higher taxa abundance between samples with low (40.5 –
were found (Table 3), displaying differences between the 1.0 mg l
1) and moderate oxygen concentrations (41.0
geographical regions. In the Pfälzerwald Mountains, – 2.0 mg l
1) were significant (U-test: p ¼ 0:009). In
stygobites were the most abundant and frequent group contrast, no differences were found between the samples
with Diacyclops languidoides c.f. occurring in every bore. with low and very low oxygen concentrations and
Troglochaetus beranecki and Antrobathynella stammeri between the moderately and highly oxygenated samples
were rare, but found exclusively in this area. Niphargus (U-test: p40.3). It seems that DO concentrations below
aquilex also occurred predominantly in the Pfälzerwald 1.0 or 0.5 mg l
1 are critical for most groundwater
Mountains with one single specimen in the Haardtrand. metazoans.
Of the stygoxenes and ubiquists only Paracyclops On a landscape scale (all samples of all geographical
fimbriatus and Diacyclops bisetosus were found fre- regions) as well as on the scale of the different regions,
quently in several bores. The other five species were the GW-Fauna-Index correlated best with total abun-
either rare or occurred only in one or two bores. dance and numbers of taxa, with most coefficients
Nematodes and oligochaetes were rare in most of the r40.6 (Appendix A). The correlations coefficients for
bores as were water mites, but on one sampling detritus, bacteria colony numbers and standard devia-
occasion, 1690 and 2945 specimens of Lepidoglyphus tion of temperature were weaker (r ¼ 0:3620:75) but
Table 3. Taxa-site matrix of the invertebrate fauna of the bores investigated

Bores

Taxa Pfälzerwald mountains Haardtrand Upper Rhine Plateau

KT 1 KT 2 KT 3 KT 4 HB/HZ HB 1 HB 2 HB 3 MB/HZ MB 1 MB 2 MB 3 MB 4 OW/HZ OW 1 OW 2 OW 3 OW 4

Stygobites
Troglochaetus beranecki 4 5
Antrobathynella stammeri 58
Niphargus aquilex 285 1 42 1
Diacyclops languidoides c.f. 173 174 21 3283 5 803 1 1318 16 11 44
Diacyclops languidoides 7 2 1 719 2
Graeteriella unisetigera 1 20 2
Bryocamptus echinatus 74
Bryocamptus typhlops 1 4
Bryocamptus zschokkei 36
Parastenocaris fontinalis borea 12 1887 10 1 22
Parastenocaris germanica 3 1 200 4
Parastenocaris psammica 3
Fabaeformiscandona latens 45 1
Fabaeformiscandona wegelini 89
Heterocypris reptans 1
Pseudocandona albicans 4 349

Stygoxenes and ubiquists

Gammarus fossarum 7
Diacyclops bisetosus 323 609 14 40 145 23 3371 164 12 106 9
Diacyclops crassicaudis 312 45 125
Eucyclops serrulatus 9 84
Macrocyclops albidus 10
Megacyclops viridis 35
ARTICLE IN PRESS

Paracyclops fimbriatus 78 1 5 34 2 84 194 1544 9

H.J. Hahn / Limnologica 36 (2006) 119–137

Thermocyclops dybowskyi 1
Tropocyclops prasinus 1
Attheyella crassa 74 1 97 87
Bryocamptus minutus 134 37
Bryocamptus pygmaeus 11
Canthocamptus staphylinus 1 110
Moraria brevipes 4 641
Parastenocaris brevipes 1 7
Candona candida 5
Candonopsis kingsleii 7
Cryptocandona reducta 1 1
Cryptocandona vavrai 2

Higher taxa
Nematoda 1075 10 5 4 1 2 2 13 20 27 1814 1172 30 82 38 78 440
Oligochaeta 167 14 80 2 2 278 4 126 62 5 106 1683 2 93 82 2 9
Acari 4 2 1 2 1615 2935 31 1 28 2 1 3 7

Total abundances per sample were aggregated by sums. Name of bores: /HZ ¼ hyporheic zone. KT ¼ Kolbental Valley, HB ¼ Klammtal Valley, MB ¼ Modenbach Valley, OW ¼ Offenbacher
125

Wald.
 ARTICLE IN PRESS
126 H.J. Hahn / Limnologica 36 (2006) 119–137

150
300 300

Metazoans per litre

Metazoans per litre
Metazoans per litre

100
200 200

100 100 50

0 0 0
n=118 n=372 n=49 n=43 n=300 n=176 n=108 n=102 n=86 n=52 n=300

absent little much very much low (0 - 1.5) high (> 2.5) very low (0-0.5) moderate (>1.0-2.0)
moderate (> 1.5 - 2.5) low (>0.5-1.0) high (> 2.0)
Detritus SD Temp. (°C) DO (mg/l)

Fig. 3. Boxplots of total abundances vs. detritus, standard deviation of temperature (SD temperature) and oxygen concentration
(DO). Classification of detritus amounts is according to Table 1.

always positive and significant (po0:001). All significant

correlations between fauna and the abiotic factors were
ro0.3, mostly ro0.2, with the exception of the trap
distance below soil surface (Appendix A) and the depth
of the traps below groundwater table (not presented
here) which were negatively correlated with the fauna
data (r40:11
ro0:65). The correlations of both these
parameters with fauna were very similar, with no
significant differences on bore level (t-test for paired
samples: p40.35, n ¼ 14). Thus, in the further analysis
and discussion, only trap distance below soil surface will
be considered.
On the scale of the bores, the correlation coefficients
of total abundance and numbers of taxa with the GW-
Fauna-Index were not different (Wilcoxon-test,
p ¼ 0:653, n ¼ 17). Thus, for a clearer presentation in
Appendix B, only the total abundance correlations were
displayed. Also, the abiotic factors were not considered
in this table, because of lacking or very weak correla-
tions with fauna. For the hyporheic sites, no significant
correlation of the total abundance was found with the
GW-Fauna-Index or any environmental parameters Fig. 4. Scatterplot of the relative abundances of stygobites and
except bacteria colony numbers at site HB/HZ. In the ubiquists/stygoxens vs. GW-Fauna-Index. Vertical lines in the
groundwater, at four bores (KT4, HB2, MB1 and MB2) plot separate the three index groups.
total abundance was not correlated with the GW-
Fauna-Index or any environmental parameters. At the
bores MB1 and MB2, groundwater tables fluctuated Rhine Plateau either no or only positive correlations
strongly with all traps drying temporarily. Bore HB2 were found.
was nearly free of fauna, and bore KT4 was strongly
influenced by unstratified groundwater from the adja- The GW-Fauna-Index
cent fractured rock aquifer. At the other 11 bores, From all environmental variables the GW-Fauna-
faunal abundance had always significant and positive Index correlated best with faunal data. Although the
correlations with the GW-Fauna-Index, while the other data were widely scattered, the proportions of stygobites
environmental parameters did not always correlate with and ubiquists/stygoxenes displayed clear patterns along
the abundance. DO was incongruent: in the Pfälzerwald the index (Fig. 4). In the lowest part of the x-axis, fauna
Mountains, it was negatively or not correlated with the was nearly absent, and in an intermediary reach of the x-
total abundance, while in the Haardtrand and the Upper axis, stygobites were the prevailing group, while its
 ARTICLE IN PRESS
H.J. Hahn / Limnologica 36 (2006) 119–137 127

upper reach was dominated by ubiquists and stygox- stygobites in the group II samples were nearly equal.
enes. According to this observation, the samples were From the group III samples, 11 of the 15 were taken
separated into three groups: group I is characterized by from the hyporheic sites (Fig. 6). The differences
index values p2.0, group II by values between 42.0 and observed between the groups were significant for both
10.0 and group III by GW-Fauna-Index values 410.0. ubiquists/stygoxenes and stygobites (U-test: po0:01),
From 60% to 85% of all group I samples, fauna with the exception of the Haardtrand for all groups and
and crustaceans, respectively, were completely absent the stygobites, and for the ubiquists/stygoxenes between
(Fig. 5), but the proportions of fauna-less samples group I and II (U-test: p40.121).
decreased via group II (13% and 27% for all taxa and Relative amount of detritus, standard deviation of
crustaceans, respectively) to 6.6% in group III samples temperature, bacteria colony numbers and DO concen-
for both total fauna and crustaceans. Both numbers tration were lowest in group I and highest in group III,
of all taxa and total abundance increased significantly while these parameters were intermediate in group II
(U-test: po0:001) and continuously from group I to (Fig. 7). The differences between the groups were
group III (Fig. 5). significant for all parameters (U-test: po0:001), with
On a landscape scale for all data, in group I samples, the exception of DO between group II and III (U-test:
crustaceans (ubiquists/stygoxenes and stygobites) were p ¼ 0:620).
rare (see above), while in group II samples, stygobites The grouping displayed above (Figs. 4–6) was
were the prevailing group accompanied by some confirmed on trap level by a discriminant analysis both
ubiquists/stygoxenes, and in group III samples, the for faunal data and for environmental parameters by
communities were dominated by ubiquists/stygoxenes 86.4% (Table 4).
(Fig. 6). The proportions of stygobites were low in Using a multiple analysis of regression to test all
group III. On the regional scale, the situation was environmental variables, total abundance and numbers
similar, but in the Pfälzerwald Mountains, group III of taxa could be best described by the GW-Fauna-Index
samples were completely lacking, and in the Haard- (Table 5). Total abundance was also influenced by COD
trand, the proportions of ubiquists/stygoxenes and and bacteria colony numbers, while the numbers of taxa

100 absent Total fauna 100 absent Crustaceans

present present

80 80
Proportion (%)

Proportion (%)

60 60

40 40

20 20

0 0
n: 228 306 15 n: 228 306 15

Group I Group II Group III Group I Group II Group III

GW-Fauna-Index GW-Fauna-Index

300 Total abundance 12 Number of taxa

10
Individuals per litre

Number of taxa

200 8

100 4

0 0
n: 228 306 15 n: 228 306 15

Group I Group II Group III Group I Group II Group III

GW-Fauna-Index GW-Fauna-Index

Fig. 5. Faunal characteristics of the three GW-Fauna-Index groups: frequency of total fauna and crustaceans, total abundance and
taxonomic richness.
 ARTICLE IN PRESS
128 H.J. Hahn / Limnologica 36 (2006) 119–137

Stygoxenes and ubiquists Stygobites

All Data Pfälzerwald Mountains

100

Proportions (%) 75

50

25

0
n= 228 228 306 306 15 15 77 77 205 205

Haardtrand Upper Rhine Plateau

100

75
Proportions (%)

50

25

0
n= 21 21 56 56 6 6 130 130 45 45 9 9

Group I Group II Group III Group I Group II Group III

GW-Fauna-Index GW-Fauna-Index

Fig. 6. Boxplots of the relative abundances of stygobites and ubiquists/stygoxens of the three GW-Fauna-Index groups. Plots refer
to all data and to the three geographical regions investigated.

were additionally explained by NH4. Bacteria colony with Dimension 1 (Spearman rank correlation: r40.8,
numbers are a reflection of food supply. NH+ 4 and COD po0:001, n ¼ 35). A multiple analysis of regression
reflect hydrochemical particularities of the geographical explained the situation of the traps along Dimension 1
regions, especially of the Upper Rhine Plateau with its best by the GW-Fauna-Index (Table 6), while their
usually low total abundances and numbers of taxa per situation along Dimension 2 was best described by
sample (see Tables 2 and 3 and negative Betas in Table temperature, total dissolved iron and the GW-Fauna-
5). However, the effects of bacteria colony numbers, Index. Temperature and total dissolved iron reflect the
NH+ 4 and COD on total abundance and numbers of particularities of the geographical regions with increas-
taxa were weak. ing values from the Pfälzerwald Mountains to the Upper
Based on the fauna data, the MDS plot (Fig. 8A) Rhine Plateau (see Table 2). The effects of the GW-
displays two groups for the traps of the Pfälzerwald Fauna-Index on Dimension 2 found in the analysis of
mountains and those of the Upper Rhine Plateau, which regression depend in particular on the two hyporheic
are separated along Dimension 2. In contrast, but in sites (Upper Rhine Plateau: OW3/HZ and Haardtrand:
accordance with their intermediate geographic location, MB4/HZ) and on the Haardtrand traps MB4/A and
the Haardtrand traps are ordered diffusely on the plot. MB3/A, which are ordered in the uppermost part of
Along Dimension 1, traps are ordered according to their dimension 2 (Fig. 8A and B). This means, the fauna of
average GW-Fauna-Index (Fig. 8B), which is well the traps in the study area were probably mainly shaped
reflected by both the numbers of taxa and the total due to two factors: the strength of the hydrological
abundance (Fig. 8C and D). GW-Fauna-Index, numbers exchange expressed by the GW-Fauna-Index and their
of taxa and the total abundance are strongly correlated affiliation to a certain geographical region.
 ARTICLE IN PRESS
H.J. Hahn / Limnologica 36 (2006) 119–137 129

Detritus (relative amount) SD Temperature (˚C)

and climatical changes occurring along groundwater
Log (Bact. col. number +1)
pathways from the Pfälzerwald Mountains to the Upper
DO (mg /l)
Rhine Plateau. Nitrate is highest in the Pfälzerwald
Mountains, although landuse is restricted on forestry —
a result of atmospheric depositions that are linked to the
10.00 acidification of the area’s uppermost aquifers (Fiedler-
Weidmann & Hahn, 1996). Bacteria colony numbers
and the amount of detritus correlated significantly with
7.50 the standard deviation of temperature, indicating that
they originated from the surface. This is plausible for
Values

bacteria colony numbers, which were strongly correlated

5.00 with SD temperature. Surface-dwelling bacteria are
permanently transported into groundwater, but density
decreases rapidly with sub-surface distance and time
2.50 (Holben & Ostrom, 2000). Cultivation of stygobiontic
bacteria is very difficult (Griebler, 2003), and cultivating
samples on standard media at 36 1C (as done in this
0.00 study) probably supports stygoxenous types (DVWK,
n= 228 228 110 217 306 306 196 306 15 15 9 15
1988). For detritus, with its much weaker correlations
Group I Group II Group III with SD temperature, conditions seem to be different.
GW- Fauna-Index
There are probably two types of detritus in the alluvial
Fig. 7. Boxplots of selected environmental variables refering groundwater: young detritus with good nutritive qual-
to the three GW-Fauna-Index groups. The 1st, 2nd (median) ity, which is directly linked with hydrological exchange
and 3rd quartile of the raw, non-transformed bacteria colony and washed into the groundwater from the surface, and
numbers (colonies per ml) is 20, 42 and 150 for group I, 26, 103 old, pre-fossile detritus with poor nutritive quality; that
and 677 for group II, and 4950, 12 000 and 26 375 for group has been deposed for hundreds or thousands of years in
III. Bact. col. number ¼ bacteria colony number. the sediments.
No correlations were found between oxygen and SD
Table 4. Stepwise discriminant analysis to test the grouping temperature in the whole study area and in the
of the traps Haardtrand, but these parameters were negatively
correlated in the Pfälzerwald Mountains and positively
Fauna (n ¼ 44) Environmental
in the Upper Rhine Plateau, reflecting regional particu-
variables (n ¼ 44)
larities. In the Pfälzerwald Mountains, most alluvial
Percent correctly 86.4 86.4 groundwater is recharged from adjacent fractured rock
classified aquifers. This provides highly oxygenated groundwater
even after a long underground passage, while surface
Discriminant power near alluvial groundwater is influenced by the anoxic
Function 1 1.00 0.77
floodplain gley soils (Hahn, 1996). Thus, oxygen
Function 2 0.68 0.47
Levels of significance concentrations usually increase with depth in this
(functions) geographic region, while SD temperature decreases. In
F1 through F2 o0.001 o0.001 the Upper Rhine Plateau both DO and SD temperature
F2 0.001 0.009 decrease with depth. Due to the special geological
conditions, the Haardtrand bores were hydro-chemi-
(a) By fauna data on species level, and (b) By the environmental cally very different, and groundwater tables varied
variables listed in Appendix A, but GW-Fauna-Index was not included
in the analysis.
strongly with the result of high variances of ground-
water temperatures indicating a strong hydrological
exchange with surface water. As a result, bacteria colony
Discussion numbers and the amount of detritus were highest in the
Haardtrand.
Environmental variables

The physical–chemical conditions are in agreement Fauna

with the characterizations of the study area made
elsewhere (e.g. Geiger et al., 1987, 1991; Geiger, 1996; General patterns
LfW, 1996; Hahn, Bauer, & Friedrich, 2000). Hydro- The results give indications for regionally typical
chemistry and water temperature reflect the geological groundwater communities, with clear patterns (Table 3,
 ARTICLE IN PRESS
130 H.J. Hahn / Limnologica 36 (2006) 119–137

Table 5. Results of a stepwise multiple linear analysis of regression

Dependent variables Order of entry Independent variables Beta r2 p

Total abundance 1 GW-Fauna-Index 0.723 0.619 o0.001

2 COD (mg l
1)
0.322 0.669 0.001
3 Bacteria colony number 0.284 0.720 0.011
Numbers of taxa 1 GW-Fauna-Index 0.874 0.713 o0.001
2 NH4 (mg l
1)
0.194 0.750 0.020

Data of the traps were aggregated by means. Independent variables included in the analysis were: GW-Fauna-Index, depth of traps below soil
surface, temperature, pH-value, buffering capacity, EC, COD, NO-3, NH+ 3-
4 , PO4 , total dissolved iron and bacteria colony number. Those variables,
which were used to calculate the GW-Fauna-Index (DO, detritus and standard deviation of temperature) were excluded. Variables included in the
analysis were entered or removed using p-values of 0.05 and 0.10, respectively. r2 indicates the improved regression coefficient of the model after the
entry of the new variable.

MB3/B MB3/C Stress: 0,16 GW-Fauna-Index Stress: 0,16

OW3/HZ
MB3/A
OW2/A OW3/A OW4/A
MB4/HZ
OW1/A
OW4/B
MB4/A MB4/B MB4/C OW2/B
MB2/C OW4/C

HB1/A
HB1/C KT3/A
HB2/A
HB3/A KT2/B
KT1/A
KT1/B KT2/C
KT4/A KT2/A KT1/C
HB1/B
HB3/C
KT4/B HB2/HZ
HB3/B

MB1/C
Pfälzerwald Mountains MB1/B
Haardtrand
A B
Upper Rhine Plateau

Numbers of taxa Stress: 0,16 Abundance Stress: 0,16

C D

Fig. 8. (A): MDS of the faunal communities by traps. Traps were exluded when crustaceans were absent. Data were aggregated by
means. Distance-measure: Bray-Curtis dissimilarity. Different symbols mark the affiliation of a trap to a geographical region.
(B–D): Fauna MDS (A) overlayed by bubbles. Size of each bubble indicates the relative mean of GW-Fauna-Index, numbers of taxa
and total abundance at each trap, respectively.

Fig. 6) in the Pfälzerwald Mountains and in the Upper distribution patterns for the stygofauna that reflect
Rhine Plateau. The Haardtrand fauna was similar to geographical, hydrographical and geological features of
both these regions because of its geographical situation, the Southwestern German state of Baden-Württemberg.
but was very heterogeneous due to the geological and According to the physical–chemical findings and the
hydrogeological diversity. Investigations on large scale geographical situation, the fauna of the Haardtrand was
distribution of stygofauna are scant, but the findings of intermediary between the Pfälzerwald Mountains and in
Fuchs and Hahn (2004) were similar. They found the Upper Rhine Plateau. In contrast, very few, weak
 ARTICLE IN PRESS
H.J. Hahn / Limnologica 36 (2006) 119–137 131

Table 6. Results of a stepwise multiple linear analysis of regression of the MDS coordinates

Dependent variables Order of entry Independent variables Beta r2 p

Dimension 1 1 GW-Fauna-Index
0.818 0.658 o0.001

Dimension 2 1 Temperature (1C) 0.614 0.523 o0.001
2 Total dissolved iron (mg l
1) 0.382 0.664 0.002
3 GW-Fauna-Index 0.70 0.727 0.012

Data of the traps were aggregated by means. Independent variables included in the analysis: GW-Fauna-Index, depth of traps below soil surface,
temperature, pH-value, buffering capacity, EC, COD, NO-3, NH+ 3-
4 , PO4 , total dissolved iron and bacteria colony numbers. Those variables, which
were used to calculate the GW-Fauna-Index (DO, detritus and standard deviation of temperature) were excluded. Variables included in the analysis
were entered or removed using p-values of 0.05 and 0.10, respectively. r2 indicates the improved regression coefficient of the model after the entry of
the new variable.

correlations were found between fauna and hydro- factors, in particular sediment characteristics (Hahn &
chemistry. This is similar to the findings of Dumas et al. Matzke, 2005; Strayer et al., 1997; Ward et al., 1998).
(2001) and Fuchs and Hahn (2004), who investigated the Detritus and bacteria colony numbers were positively
large-scale distribution of the fauna of the Arriège correlated with faunal abundance and taxonomic
alluvial aquifer in Southwestern France, and to the work richness, but so was SD temperature. Many investiga-
of Strayer et al. (1997) in the hyporheic zone of streams tions indicate the importance of organic matter as food
in the eastern USA. base for macrofauna, particularly in surface waters (see
For all samples and for most study sites, no mean- reviews by Graca, 2001; Swan & Palmer, 2000), but few
ingful correlations were found between DO and fauna, for meiofauna and groundwater. Gibert and Deharveng
but in the Klammtal Valley (Pfälzerwald Mountains), (2002) called nutrient-rich, abundantly populated zones
oxygen was negatively correlated with faunal abundance in the groundwater ‘‘islands of lifeysurrounded by
and taxonomic richness (not presented in the result (sparsely populated) deserts’’. Hahn and Matzke (2005)
section). As mentioned above, in the Pfälzerwald found that fauna density is enriched in bores because of
Mountains, DO concentrations often increase with the accumulation of detritus. Brunke and Gonser (1999)
depth, while most fauna occured near the groundwater described a strong relationship between the ratio of
table (Hahn, 2005) – probably due to a better food particulate organic carbon to total fine particles and
supply. The incongruent results of studies dealing with density patterns of hyporheic fauna, and the vertical
meiofauna and oxygen (see Malard & Hervant, 1999), distribution of particulate organic carbon was best
were probably strongly influenced by the particularities explained by the type of hydrological exchange. In the
(e.g hydraulic, hydrogeology, soils) of the study areas. groundwater, correlations between faunal data and SD
Abundances were low, or fauna was completely absent temperature probably also reflect the dependence of
from many samples with oxygen concentrations below organic matter from hydrological exchange. The results
1.0 or 0.5 mg l
1 which seems to be a critical concentra- of a recent work (Datry et al., 2005) imply that organic
tion. Although many groundwater metazoans tolerate matter is a primary factor shaping biodiversity patterns
hypoxic conditions for some time (Malard & Hervant, in the groundwater. However, many of the studies
1999), they prefer environments with an adequate available indicate the importance of organic matter for
oxygen supply. This is probably the reason, why fauna subsurface fauna.
were not always absent from groundwater with very low
oxygen concentrations, and why their abundances The GW-Fauna-Index
increased significantly if DO concentrations were higher The GW-Fauna-Index is an attempt to quantify the
than 1.0 mg l
1. If they were absent from those well ecologically relevant conditions mainly as a result of
oxygenated sites, it is very likely that this was caused by hydrological exchange between surface and ground-
other environmental factors (e.g. food shortages). Thus, water. The index was calculated using a simple
oxygen has to be considered as a limiting factor for most algorithm and those factors, which were best correlated
meiofauna. Dreher et al. (1997) found for groundwater with fauna (detritus and SD temperature) or considered
that often became anoxic, DO was governed by the as limiting (DO). It has several weaknesses, in particular
hydrological regime and was a key factor in determining that sediment structure was not considered, and that
meiofauna communities. This means that distribution of detritus was reported semi-quantitatively, and not
interstitial meiofauna is strongly limited by the factors qualitatively, in terms of C/N ratios and microbial
of scarcity of, either food supply (Strayer et al., 1997) or conditioning. C/N ratio is considered as an indicator for
DO (Hahn & Matzke, 2005; Storey & Williams, 2004), food quality (Leichtfried, 1997; Ward et al., 1998), and
though of course there are also other contributory the importance of microbial conditioned food for
 ARTICLE IN PRESS
132 H.J. Hahn / Limnologica 36 (2006) 119–137

Ranges of values are derived from the several boxplots (approximately first and third quartiles) of the result section. Bact. col. nr. (col. ml
1) ¼ Bacteria colony number (colonies ml
1). Detritus: 0
aquatic surface fauna is well known (see review of

Often absent,
Graca, 2001). Many publications emphasize the sig-

Prevailingly

Prevailingly
prevailingly

ubiquists &
stygoxenes
stygobites

stygobites
nificance of sediment structure for interstitial fauna (e.g.

Fauna
Griebler & Mösslacher, 2003; Mösslacher, 1998), but
consideration of sediment structure was not possible in
this study by methodical and financial reasons. These
weaknesses are probably the reason why fauna was

Total abund.
scattered relatively widely along the Index. On the other

(Ind. l
1)
hand, SD temperature in the index-formula, as a quasi-

450
sum parameter, seems to be corrective for the lacking

o3

50
information. However, of all environmental variables
and on all spatial scales, the GW-Fauna-Index was best

DO (mg l
1) Numbers of

correlated with total abundance and numbers of taxa,
allowing for a reasonable classification of groundwater
habitats.

1–4
taxa

43
o1
Several authors emphasize the influence of depth
below soil surface on fauna, with total abundances and
diversity decreasing with depth (e.g. Dôle-Olivier &
Marmonier, 1992; Dumas et al., 2001; Griebler &

o3.0
Mösslacher, 2003; Hahn, 2005; Hakenkamp & Palmer,

42

43
2000; Mösslacher, 1998; Strayer, 1994), but as the
regression analysis shows (Table 5), faunal data were

Characterization of the three general types of groundwater habitats identified in the study area

1 – 3.0 1C
SD temp.

43.0 1C
not explained by depth, but by the GW-Fauna-Index.

o2.0
Vertical distribution of fauna is not a question of

(1C)
absolute depth, but of the hydrological exchange, which
depends strongly on site and regional particularities.
Mösslacher (2003) argue that gradients in abundance
Bact. col. nr.

are strongly correlated with the energy supply of

100 – 1000
(col. ml
1)

groundwater systems, in terms of food and oxygen.

Much to 41000
Absent or o150

Datry et al. (2005) suggest that ‘‘depth penetration

within the sediment was limited by food supply, whereas
colonization of the upper nutrient-rich layers by
very much
Detritus

hypogean taxa was restricted by interference competi-

Little

tion’’ Poor energetic resources are probably the reason,

little

why fauna were often absent, even from shallow, well

oxygenated groundwater, and absence of fauna was
GW-Fauna-Index

strongly linked with low values of the GW-Fauna-

Index. In this sense, the GW-Fauna-Index seems to be a
42.0 – 10.0

¼ absent or little, + ¼ little, ++ ¼ much to very much.

reflection of the strength of the hydrological exchange.

The GW-Fauna-Index seems to be independent of
410.0
p2.0

regional particularities (Fig. 8). Both for all data and for
the three geographical regions investigated three groups
of groundwater habitats were distinguished. This
Moderate

classification according to three ranges of GW-Fauna-

exchange
Hydrol.

Index values is mainly based on oxygen and food

Weak

High

supply, comprising three types of habitats from oligo- to

eu-alimonic (Table 7). Oligo-alimonic type I is char-
acterized by a weak hydrological exchange, a poor food
Type II, Meso-alimonic

and oxygen supply and often an absence of fauna. At

Type I, Oligo-alimonic

Type III, Eu-alimonic

meso-alimonic type II, hydrological exchange and food

supply are moderate, oxygen concentrations are mod-
erate to high, and fauna is dominated by stygobites. Eu-
alimonic type III habitats are strongly influenced by the
Table 7.

hydrological exchange with surface water resulting in a

moderate to high oxygen and a good food supply.
Typical of the abundant and species-rich fauna are
 ARTICLE IN PRESS
H.J. Hahn / Limnologica 36 (2006) 119–137 133

ubiquists and stygoxenes. The term of ‘‘alimonic’’ [from with abundance and taxonomic richness on all spatial
lat. alimonium ¼ (food) supply] is used instead of scales and independent of the regional particularities in
‘‘trophic’’, because this latter term means per definito- the study area, allowing the distinction of three alimonic
nem the amount of inorganic nutrients that are the basis classes of groundwater habitats. However, the GW-
of primary production (Lampert & Sommer, 1993). In Fauna-Index is a long way from becoming a standar-
the groundwater, primary production is negligible. A dized technique for groundwater assessment, and in this
classification system based on the availability of organic sense it should be considered as a preliminary result.
aliments seems to be similar to the saprobic system, but Future investigations have to test, whether this ap-
saprobic systems are based on the reduction of organic proach is suitable for other areas and for sampling
matter, which affects the oxygen conditions and thereby methods other than traps. We must also determine how
the fauna communities. In contrast, groundwater com- many samples constitute the minimum requirements for
munities seem to depend directly on the organic matter. an accurate calculation of the index. In other words: a
Thus, the terms of ‘‘alimonic’’ and ‘‘alimony’’ are standard protocol for sampling is urgently required.
proposed to characterize the availability of organic However, the priority is that the gathering of detritus
aliments in the groundwater. For groundwater habitats, information has to be optimized in terms of quantitative
this means that a quantitative and qualitative record of and qualitative analyses. Perhaps, consideration of
organic matter is essential for a proper classification. sediment structure or hydraulic conductivity would also
Such information was scarcely available for the calcula- improve the accuracy of the GW-Fauna-Index. The
tion of the GW-Fauna-Index, and the problems result- alimonic classification of groundwater habitats provide
ing from this became evident in the Haardtrand (Fig. 6). important information for water supply and allow for a
In Haardtrand Group II samples, relative abundance of quality monitoring and an ecological assessment of
stygoxenes and ubiquists was higher than in the other aquifers. With respect to groundwater communities, the
areas, an observation that might be linked with a better problem of ‘‘bad predictability’’ could be minimized,
quality of detritus in the Haardtrand groundwater. All and such an index could become a base for an improved
bores there, with the exception of MB1, were obviously comparability of ecological groundwater studies.
influenced by surface water. After a heavy spate in
February 2002, high amounts of detritus occurred in the
traps, and I suspect that this was freshly imported Acknowlegements
from the surface, thus nutritive quality was better
than in the other areas. It seems that stygoxenes and I am indebted to Prof. Dr. Eckhard Friedrich for
ubiquists need a better food supply than stygobites. supporting this study and for many helpful advices, and
Groundwater pollution caused by the input of organic I thank my colleagues from the groundwater ecological
matter leads to significant changes in fauna from working group in Landau for their engaged help and
stygobiontic to stygoxenous communities (Malard many constructive remarks. Dr. Eberhard Wurst (Ho-
et al., 1996, 1999; Sket, 1999). The alimonic and henheim) and Dr. Reinhard Gerecke (Tübingen) kindly
energetic stage of most groundwater systems depends determined Lepidoglyphus destructor (Gylcyphagidae),
on the inputs from surface, and that’s what the GW- and Dr. Claude Meisch (Luxemburg) checked the
Fauna-Index displays. Fabaeformiscandona ostracods. I am indebted to the
unknown reviewers who contributed many good advices
and suggestions to this paper. For vernacular meliora-
Conclusion tion of the draft report, I thank Noel Morris. The
laboratory of the Park Brauerei, Pirmasens, kindly
The GW-Fauna-Index is the first attempt to quantify analysed the water samples for COD. This study would
hydrological exchange processes between surface water not have been possible without the support of the
and groundwater with regard to groundwater ecology. Deutsche Bundesstiftung Umwelt (DBU), Grant no.
From all environmental variables it was best correlated 20000/001.
 ARTICLE IN PRESS
134 H.J. Hahn / Limnologica 36 (2006) 119–137

Appendix A. Spearman rank correlation matrix of total abundance and numbers of taxa with
environmental variables in the whole study area (all samples) and in the three geographical regions
(Table A1)
Table A1

All samples Pfälzerwald mountains Haardtrand Upper Rhine Plateau

Abundance No. of taxa Abundance No. of taxa Abundance No. of taxa Abundance No. of taxa

GW-Fauna- r 0.6768 0.6519 0.6571 0.6101 0.4979 0.5954 0.6435 0.6013

Index
p 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000
n 549 549 282 282 83 83 184 184
SD r 0.3970 0.3755 0.5707 0.4684 0.3994 0.4910 0.5102 0.5150
temperature
p 0.0000 0.0000 0.0000 0.0000 0.0001 0.0000 0.0000 0.0000
n 584 584 300 300 91 91 193 193
Detritus r 0.5030 0.4798 0.4587 0.4939 0.6439 0.5842 0.4905 0.4167
p 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000
n 582 582 299 299 91 91 192 192
Bacterial r 0.3960 0.3605 0.4085 0.3594 0.7467 0.6488 0.5667 0.5486
abundance
p 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000
n 341 341 190 190 60 60 91 91
DO (mg l
1) r 0.2170 0.2273
0.1230
0.0224 0.0068 0.1226 0.2344 0.1750
p 0.0000 0.0000 0.0404 0.7106 0.9519 0.2727 0.0015 0.0188
n 540 540 278 278 82 82 180 180
pH-value r
0.1579
0.1583 0.1142
0.0205 0.0622 0.0936
0.0258
0.0373
p 0.0002 0.0002 0.0558 0.7325 0.5741 0.3969 0.7218 0.6068
n 558 558 281 281 84 84 193 193
EC (mS cm
1) r
0.1968
0.1892 0.0212
0.0892
0.2815
0.4064 0.0749 0.1134
p 0.0000 0.0000 0.7153 0.1237 0.0072 0.0001 0.3109 0.1245
n 574 574 299 299 90 90 185 185
Temperature r
0.1240
0.1157 0.1078 0.1079
0.1468
0.1741 0.0271 0.0637
(1C)
p 0.0032 0.0059 0.0673 0.0670 0.1855 0.1155 0.7083 0.3786
n 565 565 289 289 83 83 193 193
COD (mg l
1) r 0.0169 0.0379 0.0637 0.0245
0.1023 0.0021 0.2817 0.3181
p 0.7000 0.3873 0.3068 0.6944 0.3576 0.9849 0.0001 0.0000
n 522 522 259 259 83 83 180 180
NO3 (mg l
1) r 0.0885 0.1211
0.0088 0.0472
0.2790
0.2117 0.1920 0.2312
p 0.0496 0.0071 0.8872 0.4437 0.0147 0.0664 0.0178 0.0042
n 493 493 265 265 76 76 152 152
NH4 (mg l
1) r
0.0201
0.0479
0.0573
0.1203
0.1295
0.1139 0.1467 0.1225
p 0.6845 0.3313 0.3799 0.0645 0.3239 0.3861 0.1160 0.1903
n 413 413 237 237 60 60 116 116
PO4 (mg l
1) r
0.2284
0.1933
0.0512
0.0549 0.0153
0.0351
0.0906
0.1393
p 0.0000 0.0001 0.4186 0.3856 0.9142 0.8047 0.3271 0.1309
n 423 423 252 252 52 52 119 119
Total Fe r
0.1762
0.1502 0.0065
0.0045 0.1764 0.0437 0.0335 0.0076
(mg l
1)
p 0.0002 0.0014 0.9190 0.9439 0.1739 0.7378 0.6929 0.9286
n 451 451 249 249 61 61 141 141
 ARTICLE IN PRESS
H.J. Hahn / Limnologica 36 (2006) 119–137 135

Table A1 (continued )

All samples Pfälzerwald mountains Haardtrand Upper Rhine Plateau

Abundance No. of taxa Abundance No. of taxa Abundance No. of taxa Abundance No. of taxa

Buffering r
0.1848
0.1684 0.1217 00.0133
0.2882
0.3530
0.0156
0.0252

capacity
(mmol l
1)
p 0.0000 0.0002 0.0504 0.8312 0.0105 0.0015 0.8479 0.7568
n 490 490 259 259 78 78 153 153
Depth below n
0.3093
0.3245
0.1361
0.1120
0.3937
0.4958
0.6362
0.6143
soil surface
p 0.0000 0.0000 0.0183 0.0526 0.0001 0.0000 0.0000 0.0000
n 584 584 300 300 91 91 193 193
Depth of water r 0.0702 0.0127 0.1245 -0.0960 -0.0916
0.1757 0.0674 0.1121
table
p 0.0948 0.7621 0.0317 0.0981 0.4280 0.1264 0.3515 0.1206
n 568 568 298 298 77 77 193 193

Appendix B. Spearman rank correlation matrix of total abundance with environmental key variables
in the bores investigated (Table B1)

Table B1

Pfälzerwald mountains Haardtrand

KT1 KT2 KT3 KT4 HB1 HB2 HB3 MB1 MB2 MB3 MB4
Abund. Abund. Abund. Abund. Abund. Abund. Abund. Abund. Abund. Abund. Abund.

GW-Fauna- r 0.658 0.520 0.489
0.150 0.322 0.002 0.741
0.064
0.486 0.512 0.742
Index
p 0.000 0.000 0.006 0.446 0.048 0.992 0.000 0.836 0.329 0.008 0.000
n 45 45 30 28 38 44 39 13 6 26 36
SD r 0.732 0.836 0.523 0.104 0.597
0.174 0.797
0.538 0.301 0.840
temperature
p 0.000 0.000 0.003 0.585 0.000 0.254 0.000 0.058 0.127 0.000
n 45 45 30 30 45 45 45 13 6 27 41
Detritus r 0.434 0.383 0.279 0.125 0.403 0.027 0.646 0.000
0.655 0.647 0.785
p 0.003 0.010 0.135 0.512 0.007 0.861 0.000 1.000 0.158 0.000 0.000
n 45 45 30 30 44 45 45 13 6 27 41
Bacterial r 0.307 0.377 0.478 0.152 0.134
0.160 0.571
0.532
0.600 0.729 0.802
abundance
p 0.119 0.052 0.045 0.547 0.479 0.397 0.001 0.075 0.208 0.002 0.000
n 27 27 18 18 30 30 30 12 6 15 23
DO (mg l
1) r
0.397
0.626 0.153
0.284
0.287 0.137
0.624 0.126
0.100 0.077 0.509
p 0.007 0.000 0.421 0.144 0.076 0.406 0.000 0.682 0.873 0.708 0.002
n 45 45 30 28 39 39 39 13 5 26 36

Upper Rhine Plateau Hyporheic sites

OW1 OW2 OW3 OW4 HB/HZ MB/HZ OW/HZ

Abund. Abund. Abund. Abund. Abund. Abund. Abund.

GW-Fauna-Index r 0.793 0.701 0.489 0.365
0.045 1.000
0.223

p 0.000 0.000 0.001 0.019 0.884 0.464
n 44 41 45 41 13 2 13
 ARTICLE IN PRESS
136 H.J. Hahn / Limnologica 36 (2006) 119–137

Table B1 (continued )

Upper Rhine Plateau Hyporheic sites

OW1 OW2 OW3 OW4 HB/HZ MB/HZ OW/HZ

Abund. Abund. Abund. Abund. Abund. Abund. Abund.

SD temperature r 0.713 0.662 0.616 0.270

p 0.000 0.000 0.000 0.073
n 45 45 45 45 15 4 13
Detritus r 0.813
0.009 0.321 0.230 0.347 0.775
p 0.000 0.952 0.031 0.129 0.205 0.225
n 44 45 45 45 15 4 13
Bacterial abundance r 0.708 0.678
0.018 0.284
0.642 0.600
0.393
p 0.000 0.001 0.940 0.212 0.045 0.400 0.383
n 21 21 21 21 10 4 7
DO (mg l
1) r
0.128 0.397
0.077 0.023
0.366 1.000
0.223
p 0.407 0.010 0.623 0.889 0.218 0.464
n 44 41 43 39 13 2 13

References DVWK (1988). Bedeutung biologischer Vorgänge für die

Beschaffenheit des Grundwassers. DVWK Schriften 80.
Bretschko, G., & Leichtfried, M. (1988). Distribution of Hamburg und München: Verlag Paul Parey.
organic matter and fauna in a second order Alpine gravel Einsle, U. (1993). Crustacea: Copepoda: Calanoida und
stream (Ritrodat-Lunz study area, Austria). Verhandlungen Cyclopoida. Süßwasserfauna von Mitteleuropa, 8/4-1, Stutt-
der Internationalen Vereinigung für Limnologie, 23, gart: Gustav Fischer Verlag.
1333–1339. Fiedler-Weidmann, B., & Hahn, H. J. (1996). Quellbiotopkar-
Brown, R. J., Rundle, S. D., Hutchinson, T. H., Williams, T. tierung im Buntsandstein des Pfälzerwaldes. Unpublished
D., & Jones, M. B. (2003). Small-scale detritus-invertebrate report for Ministerium für Umwelt und Forsten Rheinland-
interactions: Influence of detrital biofilm composition on Pfalz, Mainz.
development and reproduction in a meiofaunal copepod. Fuchs, A., & Hahn, H. J. (2004). Erhebung und Beschreibung
Archiv für Hydrobiologie, 157(1), 117–129. der Grundwasserfauna in Baden-Württemberg. Unpublished
Brunke, M., & Fischer, H. (1999). Hyporheic bacteria — report of Landesanstalt für Umweltschutz Baden-Württem-
relationships to environmental gradients in a prealpine berg (LfU), Karlsruhe.
stream. Archiv für Hydrobiologie, 146(2), 189–217. Geiger, M (1996). Haardt und Weinstraße — Beiträge zur
Brunke, M., & Gonser, T. (1999). Hyporheic invertebrates –
Landeskunde. Speyer: Verlag der Pfälzischen Gesellschaft
the clinal nature of interstitial communities structured by
zur Förderung der Wissenschaften.
hydrological exchange and environmental gradients. Jour-
Geiger, M., Preuss, G., & Rothenberger, K.-H. (1987). Der
nal of the North American Benthological Society, 18,
Pfälzerwald – Porträt einer Landschaft. Landau in der Pfalz:
344–362.
Verlag Pfälzische Landeskunde.
Datry, T., Malard, F., & Gibert, J (2005). Response of
Geiger, M., Preuss, G., & Rothenberger, K.-H. (1991). Der
invertebrate assemblages to increased groundwater re-
charche rates in a phreatic aquifer. Journal of the North Rhein und die Pfälzische Rheinebene. Landau in der Pfalz:
American Benthological Society, 24(3), 461–477. Verlag Pfälzische Landeskunde.
Dole-Olivier, M.-J., & Marmonier, P. (1992). Patch distribu- Gibert, J. (2001). Basic attributes of groundwater ecosystems.
tion of interstitial communities: Prevailing factors. Fresh- In C. Griebler, D. Danielopol, J. Gibert, H. P. Nachtnebel,
water Biology, 27, 177–191. & J. Notenboom (Eds.), Groundwater ecology, a tool for
Dreher, J. E., Pospisil, P., & Danielopol, D. L. (1997). The role management of water resources (pp. 39–52). Luxemburg:
of hydrology in defining a groundwater ecosystem. In J. Office for Official Publications of the European commu-
Gibert, J. Matthieu, & F. Fournie (Eds.), Groundwater/ nity.
surface water ecotones (pp. 119–126). Cambridge: Cam- Gibert, J., & Deharveng, L. (2002). Subterranean ecosystems:
bridge University Press. A truncated functional Biodiversity. BioScience, 52(6),
Dumas, P., Bou, C., & Gibert, J. (2001). Groundwater 473–481.
Macorcrustaceans as Natural Indicators of the Ariège Graca, M. A. S. (2001). The role of invertebrates on leaf litter
Alluvial Aquifer. International Review of Hydrobiology, decomposition in streams – a review. International Revue
86(6), 619–633. der gesamten Hydrobiologie, 86(4–5), 383–393.
 ARTICLE IN PRESS
H.J. Hahn / Limnologica 36 (2006) 119–137 137

Griebler, C (2003). Mikroorganismen in Grundwasserökosys- Malard, F., Ferreira, D., Dolédec, S., & Ward, J. V. (2003).
temen. In C. Griebler, & F. Mösslacher (Eds.), Grundwas- Influence of groundwater upwelling on the distribution of
serökologie (pp. 75–158). Wien: UTB-Facultas Verlag. the hyporheos in a headwater river flood plain. Archiv für
Griebler, C., & Mösslacher, F (2003). Grundwasser – eine Hydrobiologie, 157(1), 86–116.
ökosystemare Betrachtung. In C. Griebler, & F. Mössla- Malard, F., Mathieu, J., Reygrobellet, J.-L., & Lafont, M.
cher (Eds.), Grundwasserökologie (pp. 253–310). Wien: (1996). Biomonitoring groundwater contamination: Appli-
UTB-Facultas Verlag. cation to a karst area in Southern France. Aquatic Science,
Hahn, H. J. (1996). Die Ökologie der Sedimente eines 58(2), 159–187.
Buntsandsteinbaches im Pfälzerwald—unter besonderer Malard, F., Mathieu, J., Reygrobellet, J.-L., & Lafont, M.
Berücksichtigung der Ostracoden und Harpacticiden. (1999). Groundwater contamination and ecological mon-
Marburg: Edition Wissenschaft, Reihe Biologie Bd. 62, itoring in a mediterranean karst ecosystem in southern
Tectum Verlag. France. Hydrobiologia, 58(2), 158–187.
Hahn, H. J. (2002). Developing regional models of stygofaunal Malard, F., Reygrobellet, J.-L., Mathieu, J., & Lafont, M.
communities at a landscape scale for the assessment of (1994). The use of invertebrate communities to describe
groundwater conditions – the naturraum concept. In groundwater flow and contaminant transport in a fractured
International association of hydrogeologists: International rock aquifer. Archiv für Hydrobiologie, 131, 93–110.
groundwater conference ‘‘balancing the groundwater bud- Meisch, C. (2000): Freshwater Ostracoda of Western and
get’’. Darwin NT, Australia. Central Europe. Süßwasserfauna von Mitteleuropa, 8/3,
Hahn, H. J. (2004). Grundwasser, ein bisher verkannter Heidelberg: Spektrum Akademischer Verlag.
Lebensraum. In J. Ott (Ed.), Biodiversität im Biosphärenre- Mösslacher, F. (1998). Subsurface-dwelling crustaceans as
servat Naturpark Pfälzerwald (pp. 66–78). Mainz: Bund für indicators of hydrological conditions, oxygen concentra-
Umwelt und Naturschutz Deutschland, Landesverband tions and sediment structure in an alluvial aquifer.
Rheinland-Pfalz e.V. Internationale Revue der gesamten Hydrobiologie, 83,
Hahn, H. J. (2005). Unbaited traps – a new method of 349–364.
sampling stygofauna. Limnologica, 35(4), 248–261. Mösslacher, F (2003). Evolution, Adaption und Verbreitung.
Hahn, H. J., & Friedrich, E. (1999). Brauchen wir ein In C. Griebler, & F. Mösslacher (Eds.), Grundwasseröko-
faunistisch begründetes Grundwassermonitoring und was logie (pp. 253–310). Wien: UTB-Facultas Verlag.
kann es leisten? Grundwasser, 4, 147–154. Plénet, S., Gibert, J., & Marmonier, P. (1995). Biotic and
Hahn, H. J., & Matzke, D. (2005). A comparision of abiotic interactions between surface and interstitial systems
stygofauna communities inside and outside groundwater in rivers. Ecography, 18, 296–309.
bores. Limnologica, 35(1–2), 31–44. Rumm, P. (1999). Untersuchungen zum Abbau partikulärer
Hahn, H. J., Bauer, A., & Friedrich, E. (2000). Wasser im organischer Substanzen in einem Langsamsandfilter durch
Biosphärenreservat Naturpark Pfälzerwald. Landau: Institut Metazoen am Beispiel von Niphargus fontanus BATE, 1859
für Regionale Umweltforschung und Umweltbildung. (Amphipoda, Crustacea). Unpublished Ph.D. thesis, Carl
Hakenkamp, C. C., & Palmer, M. A. (2000). The ecology of von Ossietzky Universität Oldenburg, Oldenburg.
hyporheic meiofauna. In J. B. Jones, & P. J. Mulholland Schönborn, W. (2003). Lehrbuch der Limologie. Stuttgart:
(Eds.), Streams and Ground Waters (pp. 307–336). San Schweitzerbart’sche Verlagsbuchhandlung (Nägele u. Ober-
Diego: Academic Press. miller).
Holben, W. E., & Ostrom, P. H. (2000). Monitoring bacterial Sket, B. (1999). The nature of biodiversity in hypogean waters
transport by stable isotopes enrichment of cells. Applied and how it is endangered. Biodiversity and Conservation, 8,
Environmental Microbiology, 66, 4935–4939. 1319–1338.
Janetzky, W., Enderle, R., & Noodt, W. (1996). Crustacea, Strayer, D. L. (1994). Limits to Biological Distributions. In
Copepoda; Gelyelloida und Harpacticoida. Süßwasserfauna J. Gibert, D. Danielopol, & J. A. Stanford (Eds.), Ground-
von Mitteleuropa, 8/4-2, Stuttgart: G. Fischer Verlag. water Ecology (pp. 287–310). San Diego: Academic Press
Lampert, W., & Sommer, U. (1993). Limnoökologie. Stuttgart: Inc. USA.
Georg Thieme Verlag. Storey, R. G., & Williams, D. D. (2004). Spatial responses of
Leichtfried, M. (1997). Bedsediments: Protein and POM hyporheic invertebrates to seasonal changes in environ-
content (RITRODAT-Lunz study area Austria). In mental parameters. Freshwater Biology, 49, 1468–1486.
J. Gibert, J. Mathieus, & F. Fournier (Eds.), Groundwater Strayer, D. L., May, S. E., Nielsen, P., Wolheim, W., &
and surface water ecotones – biological and hydrological Hausam, S (1997). Oxygen, organic matter and sediment
interactions and management options UNESCO Interna- granulometry as control on hyporheic animal communities.
tional Hydrology Series (pp. 75–81). Cambridge: Cambridge Archiv für Hydrobiologie, 140, 131–144.
University Press. Swan, C. M., & Palmer, M. A. (2000). What drives small-scale
LfW (1996). Synoptik der Grundwasserverhältnisse in Rhein- spatial patterns in lotic meiofauna communities? Fresh-
land-Pfalz. Jahresreihe 1971 – 1996. unpublished. report, water Biology, 44, 109–121.
Landesamt für Wasserwirtschaft Rheinland-Pfalz, Mainz. Ward, J. V., Bretschko, G., Brunke, M., Danielopol, D.,
Malard, F., & Hervant, F. (1999). Oxygen supply and the Gibert, J., Gonser, T., & Hildrew, A. G. (1998). The
adaptations of animals in groundwater. Freshwater Biology, boundaries of the river systems: The metazoan perspective.
41, 1–30. Freshwater Biology, 40, 531–569.