Seediscussions,stats,andauthorprofilesforthispublicationat:https://www.researchgate.net/publication/229681896

DiatomsasindicatorsofstreamqualityintheKathmanduValleyandMiddleHillsofNepalandIndia

ARTICLEinFRESHWATERBIOLOGY·OCTOBER2003

ImpactFactor:2.74·DOI:10.1046/j.1365-2427.2003.01138.x

CITATIONS

73

READS

151

6AUTHORS,INCLUDING:

IngridJüttner

NationalMuseumWales

61PUBLICATIONS1,179CITATIONS

SEEPROFILE

SteveJ.Ormerod

CardiffUniversity

284PUBLICATIONS8,849CITATIONS

SEEPROFILE

EileenJCox

NaturalHistoryMuseum,London

125PUBLICATIONS1,985CITATIONS

SEEPROFILE

Allin-textreferencesunderlinedinbluearelinkedtopublicationsonResearchGate,

lettingyouaccessandreadthemimmediately.

Availablefrom:SteveJ.Ormerod

Retrievedon:03February2016

APPLIED ISSUES

Diatoms as indicators of stream quality in theKathmandu Valley and Middle Hills of Nepal and India

INGRID JUTTNER* , †, SUBODH SHARMA ‡, BED MANI DAHAL ‡, S . J . ORMEROD§ ,

P. JAMES CHIMONIDES– AND EILEEN J. COX*

*Department of Botany, The Natural History Museum, London, U.K.

†Department of Biodiversity and Systematic Biology, National Museums & Galleries of Wales, Cathays Park, Cardiff, U.K.

‡Department of Environmental and Biological Sciences, Kathmandu University, Dhulikhel, Kathmandu, Nepal

§Catchment Research Group, School of Biosciences, Cardiff University, Cardiff, U.K.

–Department of Zoology, The Natural History Museum, London, U.K.

SUMMARY

1. Diatoms are recognised as indicators in temperate streams, but only recently have

assessments begun of their value in indicating stream quality in the tropics and sub-

tropics. Here, we extend previous studies by assessing stream diatom assemblages in

relation to water quality and habitat character in the Kathmandu Valley, and in the Middle

Hills of Nepal and northern India. We also assessed whether the U.K. Trophic Diatom

Index (TDI) was sufficiently portable to reveal pollution in Himalayan rivers. In the more

urbanised and highly agricultural Kathmandu Valley, we compared diatom response to

water quality classes indicated by a local invertebrate index, the Nepalese Biotic Score

(NEPBIOS).

2. Thirty and 53 streams in the Kathmandu Valley (2000) and Middle Hills (1994–96),

respectively, were sampled in October and November during stable flows following the

monsoon. Diatoms were collected in riffles, water samples taken for chemical analysis, and

habitat character of the stream channel, bank and catchment assessed using river habitat

surveys. In the Kathmandu Valley, macroinvertebrates were collected by kick-sampling.

3. In total, 113 diatom taxa were found in the Kathmandu Valley streams and 106 in the

Middle Hills. Of 168 taxa recorded, 62 occurred only in the Kathmandu Valley, 56 only in

the Middle Hills and 50 were common to both areas. Most taxa found only in the

Kathmandu Valley belonged to the genus Navicula while most taxa confined to the Middle

Hills were Achnanthes, Fragilaria and Gomphonema.

4. In the Kathmandu Valley, richness and diversity increased significantly with K, Cl, SO4

and NO3, but declined significantly with Al, Fe, surfactants and phenols. Richness here

also varied with habitat structure, being lowest in fast flowing, shaded streams with coarse

substrata in forested catchments. In all streams combined, richness increased significantly

with Si, Na and PO4, but declined significantly with increasing pH, Ca and Mg.

5. Diatom assemblage composition in the Kathmandu Valley strongly reflected water

chemistry as revealed by cations (K, Na, Mg, Ca), anions (Cl, SO4), nutrients (NO3, PO4, Si),

and also substratum composition, flow character and catchment land use. The commonest

taxa in base-poor forested catchments were Achnanthes siamlinearis, A. subhudsonis,

A. undata and an unidentified Gomphonema species; Cocconeis placentula and Navicula

minima in agricultural catchments; and Mayamaea atomus var. alcimonica, M. atomus var.

Correspondence: Ingrid Juttner, Department of Biodiversity and Systematic Biology, National Museums & Galleries of Wales, Cathays

Park, Cardiff CF10 3NP, U.K. E-mail: ingrid.juettner@nmgw.ac.uk

Freshwater Biology (2003) 48, 2065–2084

� 2003 Blackwell Publishing Ltd 2065

permitis, and Nitzschia palea at polluted sites near settlements. Diatom assemblages in none-

agricultural catchments of the Kathmandu Valley and Middle Hills were similar, but they

contrasted strongly between urban or agricultural catchments of the Kathmandu Valley

and the less intensively farmed catchments of the Middle Hills.

6. In keeping with variations in assemblage composition, most streams in the Kathmandu

Valley had higher TDI values (33–87, median ¼ 64) and more pollution tolerant taxa

(0–78%, median ¼ 16) than streams in the Middle Hills (25–82, median 45, 0–26%,

median ¼ 2). TDI values correlated significantly with measured PO4, Si, and Na

concentrations in the Kathmandu Valley, and with Si and Na concentrations in the Middle

Hills. There was some consistency between water quality classes revealed by NEPBIOS

and diatoms, but also some contrast. Water quality class I–II sites had lower TDI values

and were less species rich than water quality II sites, however, there were no significant

differences in detrended correspondence analysis (DCA) assemblage scores and relative

abundances of pollution tolerant taxa between NEPBIOS classes.

7. While diatoms in the Middle Hills indicate unpolluted or only mildly enriched

conditions, they reveal pronounced eutrophication and organic pollution in the densely

populated Kathmandu Valley. In addition, diatoms appear to respond to altered habitats

in rural agricultural and urban areas. As demands on water resources in this region are

likely to increase, we advocate the continued development of diatoms as indicators using

methods based on what appear to be consistent responses in the TDI between Europe and

the Himalaya.

Keywords: diatom diversity, Himalaya, monitoring, streams, water quality

Introduction

As a result of their importance as primary producers

in freshwater ecosystems and their rapid response to

environmental change (Stoermer & Smol, 1999),

diatoms have long been used to assess ecological

conditions and monitor environmental change in

streams and rivers of Europe, North America,

Australia, New Zealand and Japan (Chessman et al.,

1999; Prygiel, Whitton & Bukowska, 1999; Stevenson

& Pan, 1999; Hill et al., 2000; Potapova & Charles,

2002). Indices have been developed to monitor

eutrophication (Descy & Coste, 1990; Kelly & Whitton,

1995; Coring, Hamm & Hofmann, 1999), organic

pollution (Watanabe et al., 1986) and human distur-

bance (Fore & Grafe, 2002), and are now widely

applied during routine water quality surveys. By

contrast, despite the changing water quality of surface

waters, there are comparatively few studies using

diatoms as indicators of pollution in the agricultural

and densely populated regions of the sub-tropics

and tropics (Nather Khan, 1991; Lobo et al., 1996;

Silva-Benavides, 1996; Michels, 1998a,b; Gomez &

Licursi, 2001). The possible benefits for remote and

economically poor regions such as the Himalayan

Middle Hills (MH) have not yet been evaluated. This

partly reflects a dearth of taxonomic knowledge and

focused research, with only a few ecological studies

linking diatom biodiversity to environmental char-

acter (Ormerod et al., 1994; Juttner, Rothfritz &

Ormerod, 1996; Nautiyal, Nautiyal & Singh, 1996a,b;

Badoni et al., 1997; Rothfritz et al., 1997; Nautiyal

et al., 1998; Nautiyal & Nautiyal, 1999; Juttner, Cox &

Ormerod, 2000; Juttner & Cox, 2001; Nautiyal &

Nautiyal, 2002).

Much of the lower Himalaya, including the MH and

the Kathmandu Valley (KV), have been subject to

impacts from changing land-use and the expansion of

settlements (Jha, 1992; Donner, 1994). These hill

regions at c. 500–2000 m a.s.l. between the Gangetic

plains and the high Himalaya form the most densely

populated area of the Himalayan range. Growing

demands on water resources from agriculture, waste

disposal and industry have resulted in eutrophication,

organic pollution and siltation of surface waters (Ives

& Messerli, 1989; Ormerod & Juttner, 1998; Rai &

Sharma, 1998). Catchment-scale impacts on rivers are

widely believed to have followed forest clearance,

2066 I. Juttner et al.

� 2003 Blackwell Publishing Ltd, Freshwater Biology, 48, 2065–2084

erosion and agricultural intensification, including

the use of fertilisers and pesticides (Chapman &

Thompson, 1995), but quantitative evidence is difficult

to find (Collins & Jenkins, 1996; Collins, Jenkins &

Sloan, 1998; Manel, Buckton & Ormerod, 2000). In

relatively lowland areas, such as the KV, the disposal of

untreated domestic and industrial wastewaters has

affected stream water quality and in some cases, waters

are heavily polluted (Karn & Harada, 2001). While the

causes of the deterioration in water quality are well

known and the impact on biodiversity has been

demonstrated (Sharma, 1996; Ormerod & Juttner,

1998; Pandit, 1999), regular monitoring schemes are

still being developed and only include methods based

on macroinvertebrates (Sharma & Moog, 1996).

Here, we investigate diatoms from streams in

variously polluted catchments of the densely popu-

lated KV, and forested and agricultural catchments at

similar altitudes in remote areas of the MH in Nepal

and north-west India. We assessed stream character,

water chemistry and habitat structure in two highly

contrasting regions in order to (i) describe correlates

with diatom richness and diversity, (ii) investigate

variations in diatom assemblage composition, (iii)

examine whether a trophic diatom index (TDI)

developed in the U.K. could detect eutrophication in

Nepalese streams and (iv) compare diatom response

to water quality as shown by a locally derived inver-

tebrate index, the Nepalese Biotic Score (NEPBIOS)

(Sharma & Moog, 1996).

Study areas

The study areas are of contrasting character with

respect to population density, intensity of agricultural

land use, industrial activities and discharge of sewage

into running waters. In Nepal nearly half of the

population (>23 million) lives in the MH, including

the Kathmandu and Pokhara Valleys, but population

density varies substantially and generally increases

from west to east. In the KV densities are much higher

(479–1343 people km)2) compared with the study

areas in the MH (4–91 people km)2; Central Bureau

of Statistics, 1987). Agriculture in the MH focuses on

maize (70%) and wheat (49%), with the highest

production per hectare in the KV particularly because

of mineral fertilisers, pesticides and irrigation

(Donner, 1994). However, since the 1960s agricul-

tural areas have declined in the KV as a result of

urbanisation. The majority of the industry is located in

the KV and the Nepalese lowlands, mostly small

companies engaged in the manufacture of food,

textiles, carpets, wood, printing products and building

materials, such as bricks and metals. Industrial, com-

munal and domestic sewage is released into surface

waters without any or sufficient treatment, resulting in

poor stream water quality, particularly in city areas of

the KV (Sharma & Moog, 1996; Sharma, 1996).

In both areas streams were located in base-rich

and base-poor catchments, underlain by limestones,

sandstones, shales and phyllites, or metamorphic

rocks comprising gneisses, granites, quartzites and

marbles. Middle Hill catchments were located

between 79�–88� east and 27�–31� north. In north-

west India, western Nepal and central Nepal they

consist mainly of base-rich rocks whereas those in

eastern Nepal are base-poor. In the KV (85�12’–

85�30’E, 27�35’–27�47’N) base-poor catchments are

located in the northern part of the Valley, and base-

rich catchments mostly in the south.

Deciduous and mixed forests in the MH and the KV

between 1000 and 2000 m a.s.l. are characterised by

Pinus roxburghii Sarg., Alnus nepalensis D. Don, Cast-

anopsis spp. and Quercus spp. (Shrestha, 1989). In the

vicinity of settlements forests are often degraded

because of the removal of firewood and over-grazing.

In the KV forests are largely restricted to the slopes,

with some intact and protected forests remaining in

the southern part around Phulchoki.

Methods

Field investigations and laboratory procedures

Thirty first to fourth order streams in the KV were

investigated in the postmonsoon season of October–

November 2000, and 53 first to fourth order streams in

the MH during October–November 1994, 1995 and

1996. This is a period of stable flow and stable water

chemistry used previously in Himalayan surveys

(Manel et al., 2000). In the KV, between 1295 and

1680 m a.s.l., sampling locations were selected within

the cities of Kathmandu and Bhaktapur, in agricul-

tural areas, and in protected, forested catchments

(Fig. 1a). Twenty-seven streams contained sufficient

numbers of diatoms for numerical analysis. Fifteen of

these streams drained catchments with calcareous

rocks or sediments, and 12 were in catchments

Diatoms in streams of Himalayan Middle Hills 2067

� 2003 Blackwell Publishing Ltd, Freshwater Biology, 48, 2065–2084

2068 I. Juttner et al.

� 2003 Blackwell Publishing Ltd, Freshwater Biology, 48, 2065–2084

underlain by base-poor metamorphic rocks. In the

MH, streams were located between 1000 and 2000 m

a.s.l. in areas of terraced agriculture and pasture, or

catchments with scrub and forest (Fig. 1b). Nineteen

streams were located in base-rich, and 34 in base-poor

catchments.

Conductivity and pH were measured on site using

portable metres. Water chemistry samples were fil-

tered (0.45 lm) and samples for cation analysis

acidified with nitric acid. Cations and anions were

analysed by inductively-coupled plasma atomic emis-

sion spectrometry and ion chromatography. In the

KV, NO3, NO2, NH4, PO4, surfactants and phenols

were analysed on the day of sampling using spectro-

photometry [photometer MPM 3000 and test sets

14556, N1/25 (NO3–N), N4/25 (NO2–N), 14739,14544

(NH4–N), 14848 (PO4–P), 14697 (surfactants), 14551

(phenols, WTW, Weilheim, Germany)]. The accuracy

of the measurements was verified against standard

solutions (CombiCheck 10 and 20). A modified ver-

sion of the U.K. Environment Agency’s river habitat

survey was performed to assess channel, riparian and

land-use character of the streams. Variables recorded

describe channel dimensions, substratum types, flow

character, artificial and natural channel features, bank

profiles and modifications, vegetation along the banks

and land use within 50 m of the stream. These

features were recorded at 10 ‘spot-checks’ at 20-m

intervals and in a ‘sweep-up’ assessment over a 200 m

survey reach in the MH, and only in ‘sweep-up’

assessments over a 50 m survey reach in the KV

(Ormerod et al., 1997; Raven et al., 1997).

Diatoms were qualitatively collected from at least

eight stones in riffles (areas of erosion) using tooth-

brushes and the samples fixed with formalin (c. 5%

final concentration). Slides were prepared following

standard procedures (H2O2 for oxidation). At least 500

diatom valves were counted and identified to species

or subspecies at 1000· magnification (Zeiss Axioplan,

DIC), and relative abundances calculated. Most iden-

tifications were based on Krammer & Lange-Bertalot

(1986–91); Lange-Bertalot & Krammer (1989); Round,

Crawford & Mann (1990); Krammer (1997); Reichardt

(1997) and nomenclatural revisions followed the

diatom software Omnidia (Version 3, Lecointe, Coste

& Prygiel, 1999). In some cases, however, identifica-

tion requires considerable taxonomic work (Juttner

et al., 2000). In the KV macroinvertebrates were

collected by 2 min kick-sampling (0.5 mm mesh),

fixed in 75% alcohol and identified to the lowest

taxonomic level possible. The Nepalese Biotic Score

(NEPBIOS) was calculated following Sharma & Moog

(1996) to derive water quality classes.

Numerical analysis

Diatom diversity (H’) and evenness (E) were calcula-

ted (Shannon & Weaver, 1949; log-based), and rich-

ness S defined as the number of taxa at each site with

>0.5% relative abundance. Relationships between

species richness, diversity, evenness, water chemistry

and habitat character were investigated by regression

analysis (SPSS, version 6.0, Norusis, 1994).

To parameterise major gradients in water chemistry

and habitat character, environmental features were

reduced by principal components analysis (PCA) on

the correlation matrices between individual variables

to produce synoptic variates (principal compo-

nents ¼ PCs, CANOCOCANOCO, version 4.5 for Windows, Ter

Braak & Smilauer, 2002). This technique is highly

desirable with multivariate data firstly to represent

changes between sites that are also multivariate and

secondly to reduce the risk of chance relationships

between environmental variables and diatom assem-

blage composition. Slope, channel dimensions and

chemical data, apart from pH, were log-transformed

prior to PCA. Some chemical and habitat data were

only available for either the KV or the MH. Therefore,

three separate PCAs were performed using data from

the KV or the MH only, and using a combined data set

for both areas. Habitat principal components for the

KV or the MH (in brackets) were derived by differ-

entiating between variables describing channel 26

(28), bank 13 (16) and catchment character seven (six).

We did not differentiate between channel, bank and

catchment variables for the combined dataset and

used a total of 59 measured variables to derive

principal components. To derive principal compo-

nents reflecting gradients in water chemistry we used

21 measured variables for the KV, 12 for the MH and

12 for the combined data set for both areas.

To investigate changes in assemblage composition

between the MH and the KV we performed a

Fig. 1 Location of study sites in the Kathmandu Valley (a) and

the Himalayan Middle Hills in Nepal and north-west India (b),

area codes: R ¼ Roop Kund, P ¼ Pindari, S ¼ Simikot,

D ¼ Dunai, MG ¼ Mustang, MS ¼ Manaslu, Helan ¼Helambu, Langtang, M ¼ Makalu, K ¼ Kanchenjunga.

Diatoms in streams of Himalayan Middle Hills 2069

� 2003 Blackwell Publishing Ltd, Freshwater Biology, 48, 2065–2084

detrended correspondence analysis (DCA) and

explored relationships with environmental gradients

by regression. Species optima and tolerances for

common diatom species along important chemical

variables such Na, NO3–N, Ca and conductivity were

derived by weighted averaging using WACALIBWACALIB 2.1

(Line & Birks, 1990). This technique illustrates the

chemical range over which major diatom species

would be expected to occur most commonly and is

important in interpreting individual species’ occur-

rence. To relate diatom assemblage composition to

environmental gradients in the KV a canonical cor-

respondence analysis (CCA) was performed using

species relative abundances (log-transformed), chem-

istry and habitat PCs (CANOCO, version 4.5 for

Windows, Ter Braak & Smilauer, 2002). We also

performed a DCA for the KV assemblages and

compared S, H’, E and DCA scores for different

NEPBIOS water quality classes using Analysis of

Variance (ANOVAANOVA, SPSS, version 6.0, Norusis, 1994).

The U.K. TDI was calculated following Kelly (2000),

and its performance tested by regression on measured

o-PO4 concentrations. We also compared TDI values

with Na and Si concentrations, which correlated

significantly with the same water chemistry principal

component (PC2) as PO4 in the KV. The U.K. TDI is

calculated using indicator values and weights for

species groups, rather than individual taxa, and

allowed the assignment of Himalayan taxa to groups.

Our assumption, evaluated as a key part of the data

analysis, is that species groups respond to changes in

stream quality in ways that are consistent between

Europe and the Himalaya.

Results

Chemical and habitat character in Kathmandu

Valley (KV) and Middle Hill (MH) streams

Stream chemistry differed between the MH and the

KV. Nitrate–N varied between 0.0 and 1.1 mg L)1 in

the MH and 0.0–2.2 mg L)1 in the KV with means of

0.1 (±0.2 SD) and 0.4 (±0.5) mg L)1, respectively.

Conductivity varied over roughly the same range

between areas (9–345 lS cm)1 MH, 17–302 lS cm)1

KV), but the median was much higher in the KV

(149 lS cm)1) than the MH (49 lS cm)1). Median

concentrations of Si, Na and Cl in the KV were 2.4,

3.3 and 3.7 · higher than in the MH (Si: 3.4 mg L)1

MH, 8.2 mg L)1 KV; Na: 1.4 mg L)1 MH, 4.6 mg L)1

KV; Cl: 0.3 mg L)1 MH, 1.1 mg L)1 KV) while the

median concentration for Ca was 5.1· higher (range

1.1–63.1 mg L)1 KV versus 0.8–39.0 mg L)1 MH). The

range of pH was similar in both areas (5.5–8.4 MH,

6.5–8.6 KV) with pH below 7.0 at only 4.2% (MH) and

2.7% (KV) of the sites.

From the river habitat survey data, streams were

similar in size in both areas (median: water

width ¼ 3 m for MH and 4 m for KV; bank

width ¼ 9 m for MH and 10 m for KV; water

depth ¼ 0.16 m for MH and 0.20 m for KV), but

steeper in the MH (median ¼ 10� slope for MH versus

2� for KV) where flows were greater. Rapids were

extensive at 57% of MH sites but only at 3% of KV

sites, while waterfalls or cascades occurred at 77% of

MH sites but only at 13% of KV sites. Consequently,

substrata in Middle Hill streams were coarser (mean

percentage boulders 38 ± 19% SD for MH, 9 ± 13 for

KV; percentage cobbles 28 ± 12 for MH, 15 ± 11 for

KV) and exposed boulders were extensive in 91% of

the MH streams compared with only 7% in KV

streams. Smaller substratum particles such as pebbles,

gravel and sand were on average twice as common in

KV streams, and silt was only recorded in the latter.

Land use differed between the areas with urban

settlements present at more sites in the KV (11% MH

versus 33% KV). Agricultural land use was common

in catchments in both areas (67% MH, 47% KV),

however, 65% of it was extensive along one or both

stream banks in the KV compared with only 17% in

the MH. In contrast pasture was extensive at 19% of

the sites in the MH, but present at only two sites in

the KV. Mixed or conifer forests were equally

common in both areas, but the vegetation along

the riverbanks and in the catchments was often

denser in the MH (extensive forests at 49% of sites

MH versus 30% for KV; extensive scrub 57% MH

versus 3% for KV).

Diatom species richness and diversity

Kathmandu Valley streams. In the KV, 113 diatom taxa

were found, with richness varying between two and

26 at individual sites (median ¼ 15.0, mean ¼ 16.3,

SD ¼ 7.2). Diversity H’ and evenness were low and

varied between 0.08–1.32 and 0.09–0.83, respectively.

The most species rich genera were Navicula (33),

Gomphonema (nine), Achnanthidium (nine), Achnanthes

2070 I. Juttner et al.

� 2003 Blackwell Publishing Ltd, Freshwater Biology, 48, 2065–2084

(seven), Nitzschia (seven) and Planothidium (six). The

most abundant taxa were Cocconeis placentula

Ehrenberg, N. minima Grunow, N. heimansioides

Lange-Bertalot, A. subhudsonis Hustedt and Mayamaea

atomus var. permitis (Hustedt) Lange-Bertalot.

There were marked variations in richness and

diversity between sites that reflected variations in

chemistry and habitat character (Table 1). Species

richness increased significantly with increasing con-

ductivity because of elevated K, Cl and NO3 concen-

trations (water chemistry PC1, Tables 1 and 2), but

declined significantly with higher concentrations of

Al, Fe, surfactants and phenols (water chemistry PC3).

Species richness, diversity H’ and evenness also

increased significantly with increasing Si, Na and

PO4 concentrations (water chemistry PC2), and diver-

sity and evenness with increasing Mn and NH4 (water

chemistry PC4), but all fell with increasing pH, Ca

and Mg (water chemistry PC2). Some of the strongest

correlates with richness were with stream habitat

structure. In particular, richness and diversity

declined at sites with fast flow, coarser substrata

and bedrock, compared to sites with slower flow,

pebble, gravel and silt substrata (channel PC1). Spe-

cies richness, diversity H’ and evenness were also

lower in shaded streams with forested catchments,

compared to streams in agricultural catchments (bank

PC1, land use PC1).

Middle Hill streams. In the MH, 106 taxa were found.

Richness ranged between seven and 23 (med-

ian ¼ 13.0, mean ¼ 14.2, SD ¼ 4.1), diversity H’

between 0.46 and 1.09, and evenness between 0.39

and 0.81. As with the KV, some of the most species

rich genera were Gomphonema (17), Navicula (14),

Achnanthidium (12), Nitzschia (eight), and Achnanthes

(seven), but Fragilaria (eight) was also well represen-

ted. As in the KV, the most common and abundant

taxa included C. placentula, but there were also a range

of other species, such as Achnanthidium minutissimum

(Kutzing) Czarnecki, A. biasolettianum (Grunow in

Cleve & Grunow) Round & Bukhtiyarova, Fragilaria

arcus var. recta Cleve, A. subhudsonis, Diatoma mesodon

(Ehrenberg) Kutzing, Adlafia muscora (Kociolek &

Reviers) Lange-Bertalot, N. heimansioides, Reimeria

sinuata (Gregory) Kociolek & Stoermer and Achnanthi-

dium sp. 1. Diversity and evenness were lower at sites

with higher conductivity, pH, Ca and Mg concentra-

tions (water chemistry PC1, Tables 2 and 3), but

species richness was greater at sites with higher Na

and Si concentrations (water chemistry PC2). There

was no significant relationship between species

richness, diversity H’ or evenness and NO3–N (water

chemistry PC3). Few significant relationships were

found with habitat character. Only diversity H’ and

evenness were lower at sites with artificially modified

banks [F(1,15) 5.4, P < 0.05, r2 ¼ 0.31, F(1,15) 5.8,

P < 0.05, r2 ¼ 0.32].

Comparison of diatom assemblages

in the Kathmandu Valley and Middle Hills

In total 168 taxa were found in all 80 streams in the

study, 50 taxa occurring in both the KV and MH,

while 62 taxa only occurred in the KV and 56 only in

the MH. Of taxa occurring only in the KV or the MH

(in brackets), 27 (nine) belonged to the genus Navicula,

two (eight) to Achnanthes, three (eight) to Fragilaria,

and two (10) to Gomphonema. Most of the Navicula spp.

from the KV sites were rare, and only N. rostellata

Kutzing, N. schroeteri Meister and N. subminuscula

Manguin were abundant in some streams. Achnanthes

cf. holsatica, found only in the KV, was abundant in

two streams. Taxa common or abundant only in

Middle Hill streams included Achnanthidium cf.

inconspicuum, Achnanthidium sp. 5, F. arcus var. recta,

F. capucina var. vaucheriae (Kutzing) Lange-Bertalot,

Gomphonema parvulum (Kutzing) Kutzing var. parvu-

lum f. parvulum and Gomphonema sp. 6.

Diatom assemblage composition across all sites was

significantly correlated with water chemistry and

habitat character (Fig. 2a–c, Tables 4 and 5). Changes

were most pronounced along gradients from base-

poor to base-rich streams (chemistry PC1), along

salinity and nitrate concentration gradients (chemistry

PC2), and from fast flowing streams with coarse

substrata to streams in agricultural or urban catch-

ments with bank modifications and finer substrata

(RHS PC1).

Diatom assemblages in streams with higher salinity

and nitrate concentrations, close to or within settle-

ments (KTM 1, 2, 3, 6, 26, 28, 29), and base-rich streams

in intensively farmed catchments (KTM 11–15, 21, 22,

24, 25, 28), differed most from those in Middle Hill

streams (Figs 1 & 3a). Streams close to settlements were

characterised by M. atomus var. alcimonica (Reichardt)

Reichardt in Lange-Bertalot and M. atomus var. permitis,

Nitzschia palea (Kutzing) W. Smith, N. palea var. 1 and

Diatoms in streams of Himalayan Middle Hills 2071

� 2003 Blackwell Publishing Ltd, Freshwater Biology, 48, 2065–2084

Ta

ble

1W

ater

chem

istr

yan

dri

ver

hab

itat

pri

nci

pal

com

po

nen

tsre

flec

tin

gm

ajo

ren

vir

on

men

tal

gra

die

nts

inst

ream

so

fth

eK

ath

man

du

Val

ley

and

sig

nifi

can

tv

aria

ble

s

Pri

nci

pal

com

po

nen

ts

Ch

emis

try

Ch

ann

elB

ank

Lan

du

se

PC

1P

C2

PC

3P

C4

PC

1P

C2

PC

3P

C4

PC

1P

C2

PC

1P

C2

%V

aria

nce

37.7

24.2

12.9

7.5

25.9

21.9

9.3

8.2

33.8

20.9

29.7

20.7

Sig

nifi

can

t

po

siti

ve

corr

elat

ion

KS

iA

lM

nE

xp

.b

ou

lder

sB

ank

full

wid

thG

lid

es%

San

dB

ank

roo

tsC

om

po

site

ban

k

Mix

ed

fore

st

Scr

ub

FP

O4

Fe

Br

%B

ou

lder

sW

ater

wid

thP

oo

lsB

rid

ges

Ov

erh

.

bo

ug

hs

Ov

erh

.

bo

ug

hs

Pas

ture

Cl

Na

Su

rfac

tan

tsN

H4

Rap

ids

Ban

kh

eig

ht

Veg

.

po

int

bar

s

Po

ols

Sh

adin

gR

esec

tio

ned

ban

k

SO

4P

hen

ols

Ph

eno

lsE

xp

.b

edro

ckM

arg

inw

idth

%S

ilt

Tre

esS

had

ing

NO

2W

ater

fall

sW

ater

dep

th

Sr

%C

ob

ble

sU

nv

eg.

sid

ebar

Co

nd

uct

ivit

yR

iffl

esU

nv

eg.

po

intb

ar

NO

3V

eg.

sid

eb

ars

Wat

erd

epth

Sig

nifi

can

t

neg

ativ

eco

rrel

atio

n

Ca

pH

Un

veg

.

sid

ebar

s

%G

rav

el%

Gra

vel

%P

ebb

les

Gen

tle

ban

k

Ste

epb

ank

Ag

ricu

ltu

reC

on

if.

fore

st

Br

Un

veg

.

po

intb

ars

Wo

od

yd

ebri

s%

San

dIr

rig

atio

n/

div

ers.

Ban

k

mo

win

g

Ver

tica

l

+to

eb

ank

Urb

an

Mg

Ru

ns

%S

ilt

Info

rced

ban

k

Ver

t./

un

der

cut

ban

k

pH

%P

ebb

les

Co

nd

uct

ivit

y%

Gra

vel

%S

ilt

Ex

p.,

exp

ose

d;

veg

.,v

eget

ated

;u

nv

eg.,

un

veg

etat

ed;

ov

erh

.,o

ver

han

gin

g;

ver

t.,

ver

tica

l;co

nif

.,co

nif

er;

div

ers.

,d

iver

sio

n.

2072 I. Juttner et al.

� 2003 Blackwell Publishing Ltd, Freshwater Biology, 48, 2065–2084

Sellaphora seminulum (Grunow) D.G. Mann (Fig. 3b).

M. atomus var. alcimonica had the highest weighted

averages for both Na and NO3–N (7.3 and 0.64 mg L)1).

Weighted averages for M. atomus var. permitis, Nitzschia

palea, N. palea var. 1, and Sellaphora seminulum were

also higher than for most other taxa, ranging from 4.0

to 4.8 mg L)1 for Na and 0.32 to 0.44 mg L)1 for NO3–

N. C. placentula and N. minima, the most common taxa

in agricultural areas, had lower weighted averages

for Na (2.4 and 3.4 mg L)1) and for NO3–N

(0.22–0.34 mg L)1). However, C. placentula had the

second highest weighted optimum for conductivity

(155 lS cm)1) of all taxa.

The most abundant taxa in base-rich and fast-

flowing streams in less intensively farmed or forest

catchments in the MH, plus one stream in a protected

forest of the KV (KTM 23), were A. biasolettianum,

G. parvulum var. parvulum f. parvulum, A. minutissi-

mum, A. cf. inconspicuum and D. mesodon (Fig. 3b). Of

these taxa only A. biasolettianum and G. parvulum var.

parvulum f. parvulum had high weighted averages

for Ca and conductivity (16.9, 12.2 mg L)1 and 122,

102 lS cm)1). In base-poor catchments with low to

intermediate agricultural land-use A. siamlinearis

Lange-Bertalot, A. subhudsonis and N. heimansioides,

taxa with low weighted averages for Ca and conduc-

tivity (2.4, 2.6, 2.7 mg L)1 and 27, 30, 31 lS cm)1), were

most common (Fig. 3b). Assemblages were very similar

in such streams in both areas (Fig. 3a).

Assemblage composition in the Kathmandu Valley

Canonical correspondence analysis revealed signifi-

cant relationships between diatom assemblage com-

position and environmental gradients reflecting

changes in water chemistry and habitat character

(stream channel, bank, and catchment) (Fig. 4a–c,

Table 1). The CCA axis 1 accounted for 24.6% of the

explained variation in diatom composition and was

significantly correlated with water chemistry, flow

character, substratum, stream size, slope and alti-

tude. The CCA axis 2, accounting for 11.6% of the

explained variance, reflected variations in assem-

blage composition related to changes in pH, Mn and

NH4 concentrations, but also land-use from mixed

forest and pasture to settlements and agriculture, the

degree of shading and anthropogenic modifications

of the stream banks. Sites on the right-hand side of

the ordination (Fig. 4a, b) were in base-poor catch-

ments with mixed forest and less intensive agricul-

ture in the northern part of the KV. These streams

Table 2 Summary of significant correla-

tions between species richness, diversity

and evenness, and water chemistry and

habitat principal components in streams

of the Kathmandu Valley (KV) and the

Middle Hills (MH)

F(1,25 KV,1,51 MH) P-value r2

Species richness/chemistry PC1 KV 9.3 <0.01 0.52

Species richness/chemistry PC2 KV 4.5 <0.05 0.39

Diversity H’/chemistry PC2 KV 7.0 <0.01 0.47

Evenness E/chemistry PC2 KV 6.3 <0.01 0.46

Species richness/chemistry PC3 KV 4.8 <0.05 0.40

Diversity H’/chemistry PC4 KV 6.6 <0.01 0.46

Evenness/chemistry PC4 KV 6.7 <0.01 0.46

Species richness/bank PC1 KV 24.3 <0.001 0.70

Diversity H’/bank PC1 KV 20.4 <0.001 0.67

Evenness/bank PC1 KV 10.5 <0.01 0.54

Species richness/land use PC1 KV 27.6 <0.001 0.72

Diversity H’/land use PC1 KV 10.8 <0.01 0.55

Evenness/land use PC1 KV 4.6 <0.05 0.39

Diversity H’/chemistry PC1 MH 8.4 <0.01 0.38

Evenness/chemistry PC1 MH 8.1 <0.01 0.37

Species richness/chemistry PC2 MH 4.0 <0.05 0.30

Table 3 Water chemistry principal components reflecting major

environmental gradients in the Himalayan Middle Hills and

significant variables

Principal components

Chemistry

PC1 PC2 PC3

% Variance 40.9 22.7 8.7

Significant positive correlation Ca Na NO3

Conductivity Si

SO4

Mg

K

pH

Diatoms in streams of Himalayan Middle Hills 2073

� 2003 Blackwell Publishing Ltd, Freshwater Biology, 48, 2065–2084

were fast flowing with coarser substrata, and had

lower Ca, Mg, conductivity and pH than sites

elsewhere. The most common taxa here were

A. siamlinearis, Gomphonema sp. 2, A. subhudsonis

and A. undata Meister (Fig. 4c). Streams in catch-

ments with intensive agriculture, in the lower left-

hand quadrat of the ordination plot (Fig. 4a, b), had

higher conductivity with elevated K, Cl, SO4, NO2

and NO3, and were slow flowing with pebble, gravel

and silt substrata. The most abundant taxa here were

C. placentula and N. minima (Fig. 4c). Sites in the

upper left-hand quadrate (Fig. 4a, b) had higher

conductivity and concentrations of K, Cl, SO4, NO2

and NO3, but also elevated NH4. More abundant

taxa in these streams included M. atomus var.

alcimonica and M. atomus var. permitis, N. palea

Cymbella turgidula Grunow, N. germainii Wallace

and Achnanthidium sp. 3 (Fig. 4c). These streams

were located within or close to settlements. Despite

the apparent effects on diversity, there were no

significant correlations between diatom assemblage

composition shown by the first two ordination axes

and pollution caused by metals, surfactants or

phenols.

Quality assessment: applying the U.K. Trophic

Diatom Index and the Nepalese Biotic Score

Values for the U.K. TDI ranged from 33 to 87

(median ¼ 64), with 0–78% (median ¼ 16) pollution

tolerant taxa in the KV, and from 25 to 82

(median ¼ 45), with 0–26% (median ¼ 2) pollution

tolerant taxa in the MH (Fig. 5). In the KV the stream

with the lowest TDI of 33 (KTM 23) was located in a

protected forest, while most streams in agricultural

catchments with <20% pollution tolerant taxa had

TDI values between 50 and 66. At sites with >20%

pollution tolerant taxa TDI values varied between

70 and 87, with the exception of one site with a TDI

of 66.

In the KV there was a significant correlation

between the TDI at sites with <20% pollution tolerant

taxa and measured o-PO4 concentrations (F1,13 16.1,

P < 0.01, r2 ¼ 0.74, Fig. 6). The correlation was even

stronger (F1,12 21.8, P < 0.001, r2 ¼ 0.80) when the

stream in a protected forest with a diatom assemblage

typical for clean Middle Hill sites (KTM 23) was

excluded. In Middle Hill streams o-PO4 concentra-

tions were below the detection limit, but the TDI

correlated significantly with water chemistry PC2,

representing a gradient in Na and Si concentrations

(F1,51 14.1, P < 0.001, r2 ¼ 0.46). In the KV Si, Na and

o-PO4 correlated significantly with the same water

chemistry principal component (PC2, Table 1) and

TDI values also correlated significantly with Na and Si

concentrations at sites with <20% pollution tolerant

taxa (F1,13 12.4, P < 0.01, r2 ¼ 0.70, and F1,13 12.7,

P < 0.01, r2 ¼ 0.70, respectively). In contrast to o-PO4,

however, the correlation between Na, Si and TDI in

Fig. 2 Correlation between detrended correspondence analysis

(DCA) site scores for diatom assemblages in the Kathmandu

Valley and Middle Hill streams and (a) river habitat principal

component 1, (b) water chemistry principal component 1, and

(c) water chemistry principal component 2.

2074 I. Juttner et al.

� 2003 Blackwell Publishing Ltd, Freshwater Biology, 48, 2065–2084

the KV was still significant when sites with >20%

pollution tolerant taxa were included (F1,25 17.4,

P < 0.001, r2 ¼ 0.64, and F1,25 15.3, P < 0.001,

r2 ¼ 0.62, respectively).

Using the Nepalese macroinvertebrate index NEP-

BIOS, 12 streams in the KV were classified as water

quality class I–II (slightly polluted), 13 streams as

water quality class II (moderately polluted), and two

streams as water quality class II–III (critically pol-

luted). Some diatom patterns clearly reflected this

classification. For example, TDI values varied signifi-

cantly between NEPBIOS categories (F1,23 6.7,

P < 0.05), being significantly lower in the cleaner

streams of water quality class I–II (range ¼ 33–86,

mean ¼ 58.7 ± 15.0 SD) than slightly polluted

streams in class II (range ¼ 57–87, mean ¼72.1 ± 15.0 SD). In keeping with moderate pollution,

there were significantly fewer taxa (mean ¼20.5 ± 10.0 SD) at sites in water quality class I–II than

water quality class II (mean ¼ 30.8 ± 9.1 SD, F(1,22)

6.82, P < 0.05). Achnanthes siamlinearis and Gompho-

nema sp. 2 were more common at water quality I–II

sites, and G. parvulum var. 2 occurred more frequently

at water quality II sites. However, there was no

significant difference in DCA assemblage scores

between NEPBIOS categories. In addition, the mean

relative abundances of pollution tolerant diatom taxa

varied more within NEPBIOS classes than between

them: at the class I–II sites, mean percentage contri-

bution by pollution tolerant taxa was 4.5 ± 5.2 SD, but

at three sites values reached 49–66%. At water quality

class II sites, the mean relative abundances of

pollution tolerant taxa was 36.8 ± 25.9 SD.

Discussion

Chemical and habitat character in Kathmandu

Valley and Middle Hill streams

Stream chemistry and habitat character differed

between the MH and the KV reflecting intensive

Table 4 Water chemistry and river hab-

itat principal components, reflecting major

environmental gradients in streams of the

Kathmandu Valley and the Himalayan

Middle Hills, and significant variables

Principal

components

Chemistry Habitat (RHS)

PC1 PC2 PC1 PC2

% Variance 40.8 24.8 16.8 10.2

Significant

positive correlation Conductivity Si Exp. boulders Bank width

Ca Na Waterfalls Margin width

Sr Cl Slope Unveg. sidebar

Cl NO3 Boulders Water width

Mg Rapids

K Exp. bedrock

SO4

pH

Significant

negative correlation Mg Silt Shading

Gravel Overh. boughs

Reinforced bank Trees

Bank mowing Mixed forest

Urban land use

Pebbles

Agriculture land use

Exp., exposed; unveg., unvegetated; overh., overhanging.

Table 5 Relationships between water chemistry and habitat

principal components reflecting changes in environmental con-

ditions, and DCA site scores reflecting diatom assemblage

change, in streams of the Kathmandu Valley and the Himalayan

Middle Hills

F(1,77) P-value r2

Chemistry PC1/DCA site scores, axis 1 16.1 <0.001 0.41

Chemistry PC1/DCA site scores, axis 2 62.4 <0.001 0.67

Chemistry PC2/DCA site scores, axis 1 44.8 <0.001 0.60

Chemistry PC2/DCA site scores, axis 2 31.3 <0.001 0.53

Habitat PC1/DCA site scores, axis 1 57.1 <0.001 0.66

Habitat PC1/DCA site scores, axis 2 ns

Habitat PC2/DCA site scores, axis 1 ns

Habitat PC2/DCA site scores, axis 2 ns

Diatoms in streams of Himalayan Middle Hills 2075

� 2003 Blackwell Publishing Ltd, Freshwater Biology, 48, 2065–2084

2076 I. Juttner et al.

� 2003 Blackwell Publishing Ltd, Freshwater Biology, 48, 2065–2084

agricultural land use of stream catchments in the

latter. Conductivity, concentrations of NO3, PO4, Si,

Na and Cl were higher in these streams, probably

because of the use of mineral fertilisers, as well as

higher erosion, increased weathering and evapotran-

spiration from flooded terraces (Jenkins, Sloan &

Cosby, 1995; Collins & Jenkins, 1996). Intensive

agriculture and urbanisation in the KV also led to

degradation or removal of vegetation along the

riverbanks. Substratum composition showed more

fine material in the KV reflecting higher erosion in the

catchments and sediment input into streams. In

general, all these trends match those detected more

widely across the whole Himalaya (Manel et al., 2000).

Algal assemblages, species richness, diversity

and evenness

Species richness was higher in streams of the KV than

the MH. There were more species of Navicula and

Planothidium in the KV, while in the MH more taxa

belonged to the genera Gomphonema and Fragilaria.

In contrast to other studies on polluted streams

(Kawecka, 1977; Silva-Benavides, 1996; Soininen,

2002), pollution tolerant Nitzschia species were less

common than would have been expected, whereas

Navicula species were more abundant. However, most

streams in the KV were enriched by agricultural run-

off and only a few streams were within densely

populated city areas and seriously affected by sewage.

In contrast the most common taxa in Middle Hill

streams outside the KV were characteristic for unpol-

luted streams in the Himalaya and the Alps, such as

A. biasolettianum, F. arcus var. recta and D. mesodon

(Cantonati et al., 2001; Torrisi & Dell’Uomo, 2001).

Rural agricultural streams in the MH and in the KV

outside settlements were dominated by C. placentula, a

taxon typical of meso- or eutrophic streams in agri-

cultural catchments (Van Dam, Mertens & Sinkeldam,

1994; Leland, 1995; Soininen & Niemela, 2002).

Both in the KV and Middle Hill streams less diverse

assemblages occurred in base-rich streams of higher

pH, a pattern previously observed in streams of the

Indian Himalaya (Juttner & Cox, 2001). In the KV

streams with higher concentrations of NO3, PO4, Si,

Na, K and Cl were more species rich. In the MH

richness was not related to nutrient concentrations,

probably because few streams had elevated nutrient

concentrations and the gradient was small. However,

richness was higher in streams with higher concentra-

tions of Na and Si, which indicates the influence of

agriculture (Jenkins et al., 1995). Although higher

species richness was observed in mildly enriched

streams in previous studies (Nather Khan, 1991; Lobo,

Katoh & Aruga, 1995; Juttner et al., 1996; Nautiyal

et al., 1996a,b), a decline in species richness with

increasing nutrient concentrations has been reported

in urban and rural streams in the Melbourne region,

Australia (Sonneman et al., 2001). Fewer species were

found in KV streams with elevated concentrations of

metals, phenols and surfactants, similar to patterns

found in an Argentinian stream, where species rich-

ness decreased at sites receiving effluents from

chemical, textile, leather and metal industries (Gomez,

1998), and in Idaho rivers, where total number of

diatom taxa declined at mining sites (Fore & Grafe,

2002). Species richness and diversity were also influ-

enced by flow, substratum type and land use, but only

in the KV. Fewer taxa and less diverse assemblages

were found under fast flow, on coarser substrata and

in forested catchments.

Comparison of assemblages in the Kathmandu

Valley and Middle Hill streams

Assemblage composition in streams reflected changes

in habitat and chemistry resulting from anthropogenic

impacts, such as intensive agriculture and urbanisa-

tion in the KV, as well as differences in geology (base-

rich sites in India, central, western Nepal and the

southern KV, base-poor sites in the northern part of

the Valley and Eastern Nepal). While assemblage

composition in some base-poor non-agricultural

streams in the KV was similar to base-poor streams

in Eastern Nepal, most streams in the Valley were

dominated by taxa indicating eutrophication and

organic pollution. C. placentula, the most common

taxon in agricultural streams had an intermediate

weighted optimum for NO3–N but the second highest

weighted optimum for conductivity. Taxa that were

more common near settlements, such as N. minima,

N. rostellata and N. atomus var. permitis, and N. atomus

var. alcimonica, S. seminulum, and N. palea, had higher

Fig. 3 Detrended correspondence analysis of diatom assem-

blages in streams of the Kathmandu Valley and the Middle Hills;

(a) ordination of diatom assemblages along DCA axes 1 and 2

(d KV, m MH), and (b) ordination of diatom species.

Diatoms in streams of Himalayan Middle Hills 2077

� 2003 Blackwell Publishing Ltd, Freshwater Biology, 48, 2065–2084

Fig. 4 Canonical correspondence analysis

(CCA) of diatom assemblages in

streams of the Kathmandu Valley;

(a) environmental gradients correlated

with assemblage change along CCA axes

1 and 2, (b) ordination of diatom

assemblages (numbers represent sampling

sites shown in Fig. 1a, (c) ordination of

diatom species.

2078 I. Juttner et al.

� 2003 Blackwell Publishing Ltd, Freshwater Biology, 48, 2065–2084

Fig. 5 The U.K. Trophic Diatom Index calculated for Kathmandu Valley and Middle Hill streams, and percentage of pollution

tolerant taxa. Light grey shaded areas indicate organic pollution of sampling sites, dark grey shaded areas indicate combinations

of percentage of pollution tolerant taxa and TDI values, which are unlikely to be found.

Fig. 6 Correlation between the U.K.

Trophic Diatom Index calculated for

streams in the Kathmandu Valley and

measured o-PO4 concentrations.

Diatoms in streams of Himalayan Middle Hills 2079

� 2003 Blackwell Publishing Ltd, Freshwater Biology, 48, 2065–2084

optima for NO3–N but lower optima for conductivity.

In Japanese streams, these taxa had values for toler-

ance to organic water pollution between 18.2 and 41.3

(Asai, 1995). The range for all taxa was 4.9–93.6, with

low index values indicating tolerance towards organic

pollution. N. atomus var. permitis, S. seminulum, and

N. palea were classified as saprophilous species in the

same study. C. placentula and A. subhudsonis (common

in streams with forested catchments in the KV), and

taxa typical of non-agricultural areas in the MH such

as D. mesodon, and F. arcus var. recta had high index

values (50.9 and 68.3, respectively) in Japanese

streams. This indicates that they are intolerant of

organic pollution, and were classified as saproxenous,

with the exception of D. mesodon. Cosmopolitan taxa,

such as C. placentula, N. minima, M. atomus, D. mesodon,

F. arcus var. recta, indicated similar conditions in

studies on other continents (Leland, 1995; Coring

et al., 1999). There are, however, a number of taxa that

are common in the Himalayan MH, but of limited

geographical distribution or rare in other locations,

such as N. heimansioides, N. obtecta, A. siamlinearis,

Planothidium sp. 1 and several Gomphonema taxa. More

ecological information is needed before they can be

used as indicators in this area.

Environmental gradients and assemblage composition

in the Kathmandu Valley

Canonical correspondence analysis revealed that

water chemistry, habitat structure, flow type and

land use were the most important environmental

factors for diatom assemblage composition in the KV.

Chemical principal components 1 and 2 reflected

changes because of enrichment by agriculture as well

as changes in geology between base-poor underlying

bedrock in the northern part of the KV and base-rich

sediments in the south. As would be expected

(Leland, 1995) streams in catchments with intensive

agriculture were characterised by increased concen-

trations of NO3, NO2, acidic anions (Cl, SO4), K and F

(water chemistry PC1), due to the application of

fertilisers, particularly (NH4)2 SO4, and to irrigation

and increased weathering (Jenkins et al., 1995). Phos-

phate concentrations were also higher in these

streams. However, the correlation with water chem-

istry PC1 was not significant. Diatom assemblages

differed between sites in agricultural catchments and

sites affected by sewage from settlements, as well as

sites in base-poor catchments with forests or less

intensive agriculture. In a study on benthic algae in an

agriculturally dominated landscape, Munn, Black &

Gruber (2002) identified conductivity, nutrient con-

centrations, catchment land use and flow velocity as

key environmental variables, but failed to detect any

effects of substratum type on assemblage composi-

tion. Water chemistry gradients were the strongest

correlates in our study, but changes in flow character,

substratum and stream size, shading, anthropogenic

modification of the stream bank and catchment land

use were also significantly correlated with changes in

assemblage composition along CCA axes 1 and 2.

Land use, riparian and in-stream habitat change were

correlated with nutrient enrichment and higher con-

centrations in K, Cl, SO4, Si and Na. It would be

interesting to know whether particular taxa are better

indicators for either chemical or habitat change. This

is being investigated for a data set involving over 200

Himalayan streams, and will be investigated for the

KV, once a larger data set has been compiled.

Pollution assessment

Most streams in the KV had higher TDI values and

percentage pollution tolerant species than streams in

the MH, indicating more intensive agriculture and

higher sewage input into the streams. We found a

significant relationship between the U.K. TDI and

o-PO4 concentrations in 16 agricultural and non-

agricultural streams in the KV where the relative

abundance of pollution tolerant taxa was below 20%.

Index values at organically polluted sites in the KV

were higher than at unpolluted sites but not related to

PO4. There were no significant relationships between

NO3 concentrations in streams of the KV or the MH

and the TDI. However, there were significant rela-

tionships between the TDI, Na and Si concentrations

in both areas, even when organically polluted sites in

the KV were included. Na and Si reflect impacts from

agriculture in stream catchments (Jenkins et al., 1995).

It might, therefore, be advantageous to use Na or Si

concentrations in addition to PO4 concentrations for

the development of a local index. This would allow

the inclusion of sites in less intensively farmed areas,

such as the MH, where PO4 concentrations are often

below the detection limit, but also many sites in

densely populated areas, such as the KV, which are

often affected by sewage. Kelly, Penny & Whitton

2080 I. Juttner et al.

� 2003 Blackwell Publishing Ltd, Freshwater Biology, 48, 2065–2084

(1995) and Kwandrans et al. (1998) compared the

performance of several European indices for U.K.

and Polish rivers, respectively. Both found significant

correlations between the indices, and between indices

and nutrients. However, in Poland only some indices

accurately reflected differences in water quality and

the U.K. TDI could not be used to monitor eutrophic-

ation. This suggests that the response of particular taxa

to water chemistry might vary between geographical

regions, or taxa indicating different ecological condi-

tions are being combined under a single name. Several

common species in Nepal and India are not found in

European streams, and their association with partic-

ular genera or species groups in the U.K. TDI might not

reflect their true ecological optima and tolerances. We,

therefore, suggest further surveys in the Himalayan

MH and densely populated areas such as the Kath-

mandu and Pokhara Valleys to develop a Himalayan

index based on locally collected ecological information.

At most water quality class I–II sites (NEPBIOS

macroinvertebrate index) in the KV there were fewer

pollution tolerant diatom taxa and TDI values were

lower compared with water quality II sites. Water

quality class I–II sites were also less species rich than

water quality II sites. However, water quality classes

did not reflect changes in diatom assemblage compo-

sition. Triest et al. (2001) compared indices based on

diatoms, macrophytes and macroinvertebrates and

found that primary producer and macroinvertebrate

indices correlated only weakly with each other,

reflecting changes in nutrients, chloride, dissolved

oxygen and substratum character to a different extent.

Similarly macroinvertebrates and diatoms responded

in different ways to urbanisation in streams of the

Melbourne region, Australia (Sonneman et al., 2001)

and to metals in Wales (Hirst, Juttner & Ormerod,

2002). These studies suggest that different indices

should be used as complementary tools to reflect a

variety of impacts, an approach which might also be

useful in the KV, particularly if urban polluted

streams, which receive sewage and industrial efflu-

ents, are to be monitored.

Conclusions

Our results show that diatom diversity and assem-

blage composition in Himalayan regions indicate

changes in water chemistry such as base status and

salinity, as well as eutrophication and organic pol-

lution, but also indicate changes in habitat character

related to flow type, substratum composition, bank

character and catchment land use. While assemblages

in the MH are typical of clean or mildly enriched

conditions, they indicate eutrophication and organic

pollution in most of the KV streams, with the excep-

tions of those streams in forested catchments. Assem-

blages here resemble those in Middle Hill streams with

low anthropogenic impact. Demands on water

resources and freshwater habitat changes in the

Himalaya are likely to increase in the future as the

population grows rapidly, resulting in the expansion

of settlements, continued release of untreated domestic

and industrial sewage, and the heavy use of fertilisers

and pesticides in agricultural areas. We, therefore,

advocate the continued development of diatoms as

water quality indicators using methods based on what

appear to be consistent diatom responses to major

pollutants between Europe and the Himalaya.

Acknowledgments

We would like to thank Phil Brewin and Steve

Wilkinson, Cardiff University, for taking diatom sam-

ples in the field, and Seb Buckton, Cardiff University

and Roger Wyatt, NERC Centre for Ecology &

Hydrology, for conducting the river habitat surveys.

Thanks to Roger Wyatt, Jeremy Wilkinson and Alan

Jenkins, NERC Centre for Ecology & Hydrology, Vic

Din and Gary Jones, Department of Mineralogy, The

Natural History Museum, for the water chemistry

measurements, and to Dick Johnson, NERC Centre for

Ecology & Hydrology, for organising the treks in 1994–

96. Many thanks to Heike Hirst, Cardiff University,

and Bishnu Simkhada, The Natural History Museum,

for preparation of the diatom slides. Thanks to GSF

Research Centre for Environment and Health, Institute

of Ecological Chemistry, for the loan of a Zeiss

Axioplan microscope. This research was partly funded

through a BES Small Ecological Project Grant, by the

Darwin Initiative for the Survival of Species, and

through a EU Marie Curie Fellowship to IJ.

References

Asai K. (1995) Statistic classification of epilithic diatom

species into three ecological groups relating to organic

water pollution. (1) Method with coexistence index.

Diatom, 10, 13–34.

Diatoms in streams of Himalayan Middle Hills 2081

� 2003 Blackwell Publishing Ltd, Freshwater Biology, 48, 2065–2084

Badoni K., Nautiyal R., Bhatt J.-P., Kishor B. & Nautiyal

P. (1997) Variations in the epilithic diatom community

structure due to river valley projects on the Ganga

between Rishikesh and Hardwar. Proceedings of the

Indian National Science Academy Part B Biological

Sciences, 63, 523–536.

Cantonati M., Corradini G., Juttner I. & Cox E.J. (2001)

Diatom assemblages in high mountain streams of the

Alps and the Himalaya. In: Algae in Extreme

Environments. Nova Hedwigia, Beiheft, 123, 37–61.

Central Bureau of Statistics (1987) Statistical Year Book of

Nepal. National Planning Commission, Kathmandu.

Chapman G.P. & Thompson M. (1995) Water and the

Quest for Sustainable Development in the Ganges Valley.

Global Development and the Environment Series.

Mansell, London.

Chessman B., Growns I., Curreys J. & Plunkett-Cole N.

(1999) Predicting diatom communities at the genus

level for the rapid biological assessment of rivers.

Freshwater Biology, 41, 317–331.

Collins R.P. & Jenkins A. (1996) The impact of agricultural

land use on stream chemistry in the Middle Hills of the

Himalayas, Nepal. Journal of Hydrology, 185, 71–86.

Collins R.P., Jenkins A., Sloan W.T. (1998) A GIS

framework for modelling nitrogen leaching from

agricultural areas in the Middle Hills, Nepal.

International Journal of Geographical Information Science,

12, 479–490.

Coring E., Hamm A., Hofmann G. (1999) Durchgehendes

Trophiesystem auf der Grundlage der Trophieindikation mit

Kieselalgen. DVWK Materialien, 6, Deutscher Verband

fur Wasserwirtschaft und Kulturbau e.V., Bonn.

Descy J.-P. & Coste M. (1990) Utilisation des diatomees

benthiques pour l’evaluation de la qualite des eaux courants.

Rapport Final, EC contract B-71–23, Univ. Namur-

CEMAGREFF, Bordeaux.

Donner W. (1994) Lebensraum Nepal. Eine Entwicklungs-

geographie. Mitteilungen des Instituts fur Asienkunde

Nr. 226, Hamburg.

Fore L.S. & Grafe C. (2002) Using diatoms to assess the

biological condition of large rivers in Idaho (U.S.A.).

Freshwater Biology 47, 2015–2037.

Gomez N. (1998) Use of epipelic diatoms for evaluation

of water quality in the Matanza-Riachuelo (Argentina),

a Pampean plain river. Water Research, 32, 2029–2034.

Gomez N. & Licursi M. (2001) The Pampean Diatom

Index (IDP) for assessment of rivers and streams in

Argentina. Aquatic Ecology, 35, 173–181.

Hill B.H., Herlihy A.T., Kaufmann P.R., Stevenson R.J.,

McCormick F.H., Johnson C.B. (2000) Use of

periphyton assemblage data as an index of biotic

integrity. Journal of the North American Benthological

Society, 19, 50–67.

Hirst H., Juttner I. & Ormerod S.J. (2002) Comparing the

responses of diatoms and macroinvertebrates to metals

in upland streams of Wales and Cornwall. Freshwater

Biology, 47, 1752–1765.

Ives J.D. & Messerli B. (1989) The Himalayan Dilemma.

Reconciling Development and Conservation. Routledge,

London.

Jenkins A., Sloan W.T. & Cosby B.J. (1995) Stream

chemistry in the Middle Hills and high mountains of

the Himalayas, Nepal. Journal of Hydrology, 166, 61–79.

Jha P.K. (1992) Environment & Man in Nepal. Know Nepal

Series No. 5, Craftsman Press, Bangkok.

Juttner I. & Cox E.J. (2001) Diatom communities in

streams from the Kumaon Himalaya, north-west India.

In: Proceedings of the 16th International Diatom

Symposium, (Ed. A. Economou-Amilli), pp. 237–248,

University of Athens, Greece.

Juttner I., Rothfritz H. & Ormerod S.J. (1996) Diatoms as

indicators of river quality in the Nepalese Middle

Hills with consideration of effects by habitat-specific

sampling. Freshwater Biology, 36, 475–486.

Juttner I., Cox E.J. & Ormerod S.J. (2000) New or poorly

known diatoms from Himalayan streams. Diatom

Research, 15, 237–262.

Karn S.K. & Harada H. (2001) Surface water pollution in

three urban territories of Nepal, India, and

Bangladesh. Environmental Management, 28, 483–496.

Kawecka B. (1977) Biocenosis of a high mountain stream

under the influence of tourism. 3. Attached algae

communities in the stream Rybi Potok (the High Tatra

Mts, Poland) polluted with domestic sewage. Acta

Hydrobiologica, 19, 271–292.

Kelly M.G. (2000) Identification of common benthic

diatoms in rivers. Field Studies, 9, 583–700.

Kelly M.G. & Whitton B.A. (1995) The Trophic Diatom

Index: a new index for monitoring eutrophication in

rivers. Journal of Applied Phycology, 7, 433–444.

Kelly M.G., Penny C.J. & Whitton B.A. (1995)

Comparative performance of benthic diatom indices

used to assess river water quality. Hydrobiologia, 302,

179–188.

Krammer K. (1997) Die cymbelloiden Diatomeen. Eine

Monographie der weltweit bekannten Arten, Vol. 1, 2.

J. Cramer, Berlin.

Krammer K. & Lange-Bertalot H. (1986–91).

Bacillariophyceae. In: Sußwasserflora von Mitteleuropa

(eds. H. Ettl, J. Gerloff, H. Heynig and D.

Mollenhauer), vol. 2, parts 1–4. Gustav Fischer Verlag,

Stuttgart.

Kwandrans J., Eloranta P., Kawecka B. & Wojtan K.

(1998) Use of benthic diatom communities to evaluate

water quality in rivers of southern Poland. Journal of

Applied Phycology, 10, 193–201.

2082 I. Juttner et al.

� 2003 Blackwell Publishing Ltd, Freshwater Biology, 48, 2065–2084

Lange-Bertalot H. & Krammer K. (1989) Achnanthes.

Monographie der Gattung. Bibliotheca Diatomologica,

Vol. 18. J. Cramer, Berlin.

Lecointe C., Coste M., Prygiel J. (1999) OMNIDIA

version 3: notice d’utilisation. CLCI, Monbazillac.

Leland H.V. (1995) Distribution of phytobenthos in the

Yakima River basin, Washington, in relation to

geology, land use, and other environmental factors.

Canadian Journal of Fisheries and Aquatic Sciences, 52,

1108–1129.

Line J.M. & Birks H.J.B. (1990) WACALIB v. 2.1. A

computer program to reconstruct environmental

variables from fossil assemblages by weighted

averaging. Journal of Paleolimnology, 3, 170–173.

Lobo E.A., Katoh K. & Aruga Y. (1995) Response of

epilithic diatom assemblages to water pollution in

rivers in the Tokyo Metropolitan area, Japan.

Freshwater Biology, 34, 191–204.

Lobo E.A., Callegaro V.L., Oliveira M.A., Salomoni S.E.,

Schuler S. & Asai K. (1996) Pollution tolerant diatoms

from lotic systems in the Jacuı basin, Rıo Grande do

Sul, Brazil. Iheringia, Ser. Bot., 47, 45–72.

Manel S., Buckton S.T., Ormerod S.J. (2000) Testing large-

scale hypotheses using surveys: the effects of land use

on the habitats, invertebrates and birds of Himalayan

rivers. Journal of Applied Ecology, 37, 756–770.

Michels A. (1998a) Use of diatoms (Bacillariophyceae) for

water quality assessment in two tropical streams in

Costa Rica. Revista de Biologia Tropical, 46, 143–152.

Michels A. (1998b) Effects of sewage water on diatoms

(Bacillariophyceae) and water quality in two tropical

streams of Costa Rica. Revista de Biologia Tropical, 46,

153–175.

Munn M.D., Black R.W. & Gruber S.J. (2002) Response of

benthic algae to environmental gradients in an

agriculturally dominated landscape. Journal of the

North American Benthological Society, 21, 221–237.

Nather Khan I.S.A. (1991) Effect of urban and industrial

wastes on species diversity of the diatom community in

a tropical river, Malaysia. Hydrobiologia, 224, 175–184.

Nautiyal P. & Nautiyal R. (2002) Altitudinal variations in

the relative abundance of epilithic diatoms in some

glacier and spring fed Himalayan tributaries of the

Ganga (Ganges) River system in the Garhwal region.

In: Proceedings of the 15th International Diatom

Symposium (ed. J. John), pp. 237–248, A.R.G. Gantner,

Lichtenstein.

Nautiyal P, Nautiyal R., Rawat V.S., Bhatt J.P. & Kishar B.

(1998) Stream regulation. II. Variations in the density

of phytobenthic diatoms in impounded sections of

River Ganga. In: Environment and Development. (Eds I.S.

Grover & A.K. Thukral), pp. 237–240. Scientific

Publishers, Jodhpur, India.

Nautiyal R. & Nautiyal P. (1999) Altitudinal variations in

the pennate diatom flora of the Alaknanda – Ganga

River system in the Himalayan stretch of Garhwal

region. In: Proceedings of the 14th International Diatom

Symposium (eds S. Mayama, M. Idei & I. Koizumi),

Koeltz Scientific Books, Koenigstein.

Nautiyal R., Nautiyal P. & Singh H.R. (1996a) Impact of

Sewage on the Diatom Communities of River

Alaknanda (Srinagar, Garhwal). International Journal

of Ecology and Environmental Sciences, 22, 289–296.

Nautiyal R., Nautiyal P. & Singh H.R. (1996b)

Community Structure of Coldwater Epiphytic

Diatoms in Relation to Substrate and Flow

Conditions of a Himalayan River Alaknanda. Journal

of Freshwater Biology, 8, 1–5.

Norusis M.J. (1994) SPSS for Windows, version 6.0. SPSS,

Chicago.

Ormerod S.J. & Juttner I. (1998) Catchment Sustainability

and River Biodiversity in Asia: A Case Study from

Nepal. In: The Sustainability of Tropical Catchments. (eds

D. Harper and A. Brown), pp. 187–207, John Wiley,

Chichester.

Ormerod S.J., Rundle S.D., Wilkinson S.M., Daly G.P.,

Dale K.M., Juttner I. (1994) Altitudinal trends in the

diatoms, bryophytes, macroinvertebrates and fish of a

Nepalese river system. Freshwater Biology, 32, 309–322.

Ormerod S.J., Baral H.S., Brewin P.A., Buckton S.T.,

Juttner I., Rothfritz H. & Suren A.M. (1997) River

Habitat Surveys and Biodiversity in the Nepal

Himalaya. In: Freshwater Quality: Defining the

Indefinable? (eds P.J. Boon and D.L. Howell), pp. 241–

250. The Stationary Office, Edinburgh.

Pandit A.K. (1999) Freshwater Ecosystems of the Himalaya.

The Parthenon Publishing Group, New York.

Potapova M.G. & Charles D.F. (2002) Benthic diatoms in

USA rivers: distribution along spatial and environ-

mental gradients. Journal of Biogeography, 29, 167–187.

Prygiel J., Whitton B.A. & Bukowska J. (1999) Use of Algae

for Monitoring Rivers III. Agence de l’Eau Artois-

Picardie, Douai.

Rai S.C. & Sharma E. (1998) Comparative assessment of

runoff characteristics under different land use patterns

within a Himalayan watershed. Hydrological Processes,

12, 2235–2248.

Raven P.J., Fox P., Everard M., Holmes N.T.H. & Dawson

F.H. (1997) River habitat structure: a new system for

classifying rivers according to their habitat quality. In:

Freshwater Quality: Defining the Indefinable? (eds P.J.

Boon and D.L. Howell), pp. 215–234. The Stationary

Office, Edinburgh.

Reichardt E. (1997) Taxonomische Revision des Arten-

komplexes um Gomphonema pumilum (Bacillario-

phyceae). Nova Hedwigia, 65, 99–129.

Diatoms in streams of Himalayan Middle Hills 2083

� 2003 Blackwell Publishing Ltd, Freshwater Biology, 48, 2065–2084

Rothfritz H., Juttner I., Suren A. & Ormerod S.J. (1997)

Epiphytic and epilithic diatom communities along

environmental gradients in the Nepalese Himalaya:

implications for the assessment of biodiversity and

water quality. Archiv fur Hydrobiologie, 138, 465–482.

Round F.E., Crawford R.M. & Mann D.G. (1990) The

Diatoms. Biology and Morphology of the Genera. Cam-

bridge University Press, Cambridge.

Shannon C.E. & Weaver W. (1949) The Mathematical

Theory of Communication. The University of Illinois

Press, Urbana, IL.

Sharma S. (1996) Biological assessment of water quality in the

rivers of Nepal. PhD thesis, University of Vienna,

Austria.

Sharma S. & Moog O. (1996) The applicability of biotic

indices and scores in water quality assessment of

Nepalese rivers. In: Ecohydrology of High Mountain

Areas (eds S.R. Chalise et al.) pp. 641–657, Proceedings of

the International Conference on Ecohydrology of High

Mountain Areas, ICIMOD, UNESCO, Kathmandu,

Nepal.

Shrestha B.P. (1989) Forest Plants of Nepal. Educational

Enterprise Pvt. Ltd, Lalitpur, Nepal.

Silva-Benavides A.-M. (1996) The epilithic diatom flora of

a pristine and a polluted river in Costa Rica, Central

America. Diatom Research, 11, 105–142.

Soininen J. (2002) Responses of epilithic diatom

communities to environmental gradients in some

Finnish rivers. Internationale Revue der Gesamten

Hydrobiologie, 87, 11–24.

Soininen J. & Niemela P. (2002) Inferring the phosphorus

levels of rivers from benthic diatoms using weigthed

averaging. Archiv fur Hydrobiologie, 154, 1–18.

Sonneman J.A., Walsh C.J., Breen P.F. & Sharpe A.K.

(2001) Effects of urbanization on streams of the

Melbourne region, Victoria, Australia. II. Benthic

diatom communities. Freshwater Biology, 46, 553–565.

Stevenson R.J. & Pan Y. (1999) Assessing environmental

conditions in rivers and streams with diatoms. In: The

Diatoms: Applications for the Environmental and Earth

Sciences (eds F. Stoermer, J.P. Smol), pp. 11–40,

Cambridge University Press, Cambridge.

Stoermer E.F. & Smol J.P. (1999) The Diatoms: Applications

for the Environmental and Earth Science. Cambridge

University Press, Cambridge.

Ter Braak J.F. & Smilauer P. (2002) CANOCO – Software

for Canonical Community Ordination (version 4.5).

Biometris, Wageningen and Ceske Budejovice.

Torrisi M. & Dell’Uomo A. (2001) Les diatomees benthi-

ques des parties rhithrales et potamales des cours d’eau

de l’Apennin central (Italie) et leurs significations

ecologiques. Algological Studies, 102, 35–47.

Triest L., Kaur P., Heylen S. & De Pauw N. (2001)

Comparative monitoring of diatoms, macro-

invertebrates and macrophytes in the Woluwe River

(Brussels, Belgium). Aquatic Ecology, 35, 183–194.

Van Dam H., Mertens A. & Sinkeldam J. (1994) A coded

checklist and ecological indicator values of freshwater

diatoms from the Netherlands. Netherlands Journal of

Aquatic Ecology, 28, 117–133.

Watanabe T., Asai K., Houki A., Tanaka S. & Hizuka T.

(1986) Saprophilous and eurysaprobic diatom taxa to

organic water pollution and Diatom Assemblage Index

(DAIpo). Diatom, 2, 23–73.

(Manuscript accepted 30 July 2003)

2084 I. Juttner et al.

� 2003 Blackwell Publishing Ltd, Freshwater Biology, 48, 2065–2084