Riparian zone characteristics, fluvial attributes and watershed land-use, and the utility of river otter (Lutra lutra) as indicator species, in rural Aragón, Spain

Riparian Zone Characteristics, Fluvial Attributes and Watershed Land-use, and the Utility of River Otter

(Lutra lutra)

as Indicator Species, in Rural

Aragbn, Spain.

Maria Grau L6pez

B.A. University of Victoria, 1998

A Thesis submitted in Partial Fulfillment of the Requirements for the Degree of

MASTER OF SCIENCE in the Department of Biology

W e accept this thesis as conforming

to the required standard

O Maria Carmen Grau Lopez, 2004

University of Victoria

All rights reserved. This thesis may not be reproduced in whole or in part, by photocopy or other means, without the permission of the author.

Co-Supervisor: Dr. Richard Hebda Co-Supervisor: Dr. Don S. Eastman Co-Supervisor: Dr. Asit Mazumder

ABSTRACT

Rivers in the Iberian Peninsula have been modified as a result of watershed deforestation, land conversion to agriculture, and riparian fragmentation for millennia. Recently, rivers have also been altered by construction of flood works, dams, roads, and by gravel mining. The nature and extent of these impacts varies in different river systems, and this variation provides an excellent opportunity to examine inter-relationships between riparian zone and watershed vegetative cover and fluvial attributes. As well, it provides an opportunity to assess the utility of a species, like the European otter (Lutra lutra), as indicator of the

biophysical integrity of riverine systems.

I examined five Pyrenean and pre-Pyrenean rivers in the province of Huesca, Spain. Attributes of the riparian vegetation were quantified through the interpretation of 580 airphoto strips 50m long perpendicular to the channel on each bank and field surveys. The relative abundance of otter and stream characteristics were determined in 600m

transects, with a total of 22 transects distributed among the five rivers. The proportion of forest area within a 50m wide strip was significantly and negatively correlated with channel-wetted width. Similarly in upland valley reaches, the amount of forest within the 50m wide strip was significantly and negatively correlated with the channel-wetted width to depth ratio. These findings show the importance that riparian forests have in the

protection of the banks, resulting in narrower and deeper channels. However, mature trees in the riparian strip are too scarce to supply enough large woody debris to significantly influence channel and aquatic habitat structure.

The density of otter spraints per length of bank surveyed was positively correlated to the wetted width of the channel, and the density of otter spraints per unit of wetted surface was positively associated with the proportion of residual pool habitat per channel length, and negatively correlated with sparsely vegetated areas within 50m of the active channel. These findings show the importance of pools and riparian vegetation for the river otter. Although the otter cannot be used as indicator of ecological integrity, since it may not reveal changes in fish community composition, in rivers unimpaired by drastic alterations such as dams and urban or agricultural effluents, otter density would indicate abundance of fish and, in turn, the structural condition of the aquatic habitat. In addition, since the river otter is a necessary component of the ecological integrity of the riverine ecosystem, it can play the role of a 'flagship' species because of its charismatic image.

Table of Contents Title page

...

i

...

Abstract ii

...

Table of Contents iv List of Tables

...

vii

List of Figures

...

x Acknowledgements

...

xii

_...

...

Dedication xi11 Chapter 1

.

1

.

2

.

3

.

4

.

5

.

6

.

Chapter 2

.

Introduction

...

1 The issue

...

2

The riverine ecosystem

...

4

...

Effects of watershed development on riverine systems 10 The influence of vegetation on channel morphology

...

12

...

Use of indicator species for ecosystem evaluation 15 Hypotheses

...

19

...

Approach 20 The relationship of European otter Lutra lutra L . to morphological

...

characteristics of fluvial systems of Huesca. Spain 21 Introduction

...

22

Study area

...

24

Method ... 26

Otter survey

...

26

Reach morphology and elements of fish habitat ... 28

...

Morphology type 29 Fluvial network zone

...

30

Channel planform

...

30

Biological zone

...

30

...

System disturbance and human presence 31 Statistical analysis ... 31

Results

...

31

Density of otter signs ... 32

Reach morphology. elements of fish habitat and otter signs

...

32

...

.

.

4 2.1 Differences between rivers 32 4 . 2

.

2 . Differences between reach types

...

33

Morphology type

...

33

...

Fluvial network zone 34

Channel planform

...

35

...

Biological zone 36

...

4

.

3 . Otter signs and channel attributes 36

4

.

4

.

Otter signs. river system disturbance and human presence

...

38

5

.

Discussion

...

38

5.1

.

Channel attributes and its relationship to otter occurrence

...

38

...

5

.

2

.

The use of spraints for comparison of otter abundance 45

5.3. Comparison of morphological characteristics

...

between reach types 48

...

. .

5 4 Assumptions and limitations 52

...

6

.

Conclusion 54

Tables

...

56

...

Figures 65

Chapter 3

.

The influence of vegetation type. riparian strip width. and

watershed cover on stream habitat characteristics. and on

the occurrence of European otter in rivers of Huesca. Spain

...

73

1 . Introduction

...

74

...

.

2 Study area 76

...

3

.

Methods 78

.

...

3 1

.

Selection of sites 78

3

.

2 . Drainage basin and land use

...

79

3

.

3

.

Vegetation of the riparian zone

...

79

...

3.4. Channel morphology. elements of fish habitat and otter survey 81

...

.

3 5 . Statistical analysis 84 4 . Results

...

84

...

. .

4 1 Characteristics of the drainage basins 84

...

. .

4 2 Riparian zone vegetation characteristics 85

...

Differences among rivers 85

...

Difference between reach location and channel type 86

4 . 3

.

Reach morphology ... 87

...

Differences between morphology types 87

4

.

4

.

4

.

5 . 4

.

6 . 5 . 5

.

1

.

5

.

2

.

5

.

3

.

5.4. 6 . Chapter 4

.

1

.

2

.

3

.

4 .

...

Differences between channel planform types 89

Relationship of drainage basin characteristics to

morphological features

...

89

Relationship of riparian vegetation strip characteristics to morphological features

...

90

Relationship of otter occurrence to riparian vegetation strip and drainage basin characteristics

...

91

...

Discussion 92

...

Drainage basin influence 92 The influence of riparian vegetation type on river morphology

...

94

Otter ecology and riparian vegetation

...

96

Assumptions and limitations

...

99

Conclusion

...

100

...

Tables 103

...

Figures 113

...

Conclusions 122

...

The role of the riparian vegetation in fluvial ecosystem integrity 123

...

The otter-riverine sytem relationship 131 Fluvial systems within the constraints of the watershed

...

136

Land use and historical changes to Spanish river systems -

...

considerations for the potential restoration of rivers in Huesca 139

...

Figures 144 Literature cited

...

151

vii

List of Tables

Table 1 .l. The four categories of focal species and their definitions

(Zacharias and Roff 2001 ).

...

1 6

Table 2.1. Characteristics of the five river sections examined in this study (All are located on the western side of the Ebro River watershed

in the province of Huesca).

...

56

Table 2.2. Variables measured or estimated at locations along the

transect with and without otter signs

...

57

Table 2.3. Summary of measurements and characteristics collected on the five study reaches; as from observations and

measurements taken in a 600m transect in each reach.

...

58

Table 2.4. Density of otter spraints per 100m of surveyed bank and per 100m2of wetted surface and morphological characteristics of the

...

stream, as measured in the 600 m transects of sampled reaches. 59

Table 2.5. Characteristics of the five river sections of the study-gradient, channel wetted width, active width to wetted width ratio, pool depth, wetted width to depth ratio, percent of glide-pool, riffles and rapid-cascade habitat per channel length, number of LWD accumulations and number of boulders per 100m of channel length, and mean density of

otter spraints per 100m2 of wetted surface

...

60

Table 2. 6. Morphological characteristics of two morphology types -pool/riffle and plane-bed/cascade. Pool spacing, wetted

widthldepth ratio (ww / d), active channel width to wetted

width ratio (Actw

/

ww), percent of pool-glide habitat per

channel length (PO-GL %), and number of large woody debris

per 100m of channel length (LWD #

/

100m).

...

61

Table 2. 7. Density of otter spraints per 100m2 of wetted surface,

morphological characteristics and habitat elements measured in the three fluvial network zones: upland,

upland valley and floodplain valley-percent of pool-glide habitat per length of channel; wetted widthldepth ratio, number of LWD per 100m of channel length, and number

of boulders per 1000m2 of wetted surface.

...

62

Table 2.8. Density of spraints and morphological characteristics per planform channel type: single-thread, braided channel and wandering channel;

- density of otter spraints per 1 OOm* of wetted surface, proportion of

pool-glide habitat per channel length (calculated from riffle/pool morphology sections), wettted width per depth ratio, and amount of

Table 2. 9. Density of otter spraints and morphological differences between the two biological zones rhithron and potamon-gradient, altitude, amount

of pool-glide and rapid-cascade habitat

...

64

Table 3.1. Characteristics of the five river sections examined in this study. 1 ) Drainage area measured to the lowest point of each study section; 2) Data obtained from gauging stations of the Confederaci6n Hidrografica del Ebro (CHE); 3) Garcia

Ruiz et al., 2001 ; 4) Flow in Lascellas and Ballovar, inmediately

upstream and 24 km downstream from the study section, respectively; 5) Flow in Torla and Fiscal, at the upper and the

lower limits of the study section, respectively; 6) Flow in Capella,

16 km downstream from the study section; 7) Flow in Binies,

at reach 5-6 of the study section.

...

103

Table 3.2. Land use categories and riparian vegetation classes identified on the study's drainage basins, on a 50m wide riparian corridor, and

at otter spraint and non-spraint (control) sites.

...

104

Table 3.3. In-stream features measured or estimated along the 600m

transectat locations with and without otter signs

... 105

Table 3.4. Mean proportion of land use cover types in the study reaches

sub-drainage basins within the five rivers of the study

...

106

Table 3. 5. Mean proportion of the vegetation classes within a 50m wide riparian band along both banks of the study reaches per river-MT: mature trees; YT: young trees; Treed: mature and young trees; TS: tall shrubs; Wooded: trees and tall shrubs; SS: short shrubs; SP: sparsely vegetated; AG: agricultural fields;

-

and the median width of the natural vegetation band on

each side of the channel.

...

107

Table 3.6.

Table 3. 7.

Table 3.8.

Proportion of sites with different natural vegetation band

width classes in the five study rivers. Band width was calculated from airphotos at sampling sites in both banks (N=81x2=162), up to

100m from the border of the active channel.

...

108

Morphological characteristics of two morphology types -pool/riffle and plane-bed/cascade. Active width (Actw), wetted widthldepth ratio (wwld), active channel width to wetted width ratio (Actw/ww), percent of pool-glide habitat per channel length (PO-GL %), and number of large woody

debris per 100m of channel length (LWD #/loom).

...

109

Riparian and morphological characteristics of reaches in the three fluvial network zones: upland, upland valley and floodplain

valley-proportion of forest in a 50m wide riparian band (MT & YT),

channel active width (Actw), wetted width to depth

ratio (ww / d ratio), active width to wetted width ratio (Actw/ww),

percent of pool-glide habitat per channel length (PO-GL %),

Table 3. 9. Differences in morphological characteristics per channel planform type-single-thread, braided and wandering channel- proportion of forest in a 50m wide riparian band

(MT & YT), channel active width (Actw), wettted width to depth ratio (ww/d), active width to wetted width

ratio (Actw/ww ); percent of pool-glide habitat per

channel length (PO-GL %), and amount of large woody debris

per 100 m of channel length (LWD

/

100m). 1 .Calculated on

pool / rifflemorphology reaches only. ... 1 1 1

Table 3.1 0. Spearman's Rank Correlation coefficients for percent land cover in drainage basin versus habitat and otter signs density

List of Figures

Figure 2.1 . Figure 2

.

2 . Figure 2

.

3

.

Figure 2.4. Figure 2

.

5

.

Figure 2

.

6

.

Figure 2

.

7

.

Figure 2.8. Figure 3

.

1

.

Figure 3

.

2 . Figure 3.3. Figure 3.4. Figure 3

.

5

.

Figure 3

.

6 . Figure 3 . 7

.

Figure 4.1. Figure 4 . 2

.

Figure 4.3.

Map of the five rivers of the study

...

65

Box and whisker plot of the density of otter spraints in the five river sections

...

66

Scatter plot showing the significant correlation between otter spraint densities and percent of pool-glide

...

67

Histogram of habitat type at spraint and non-spraint sites

...

68

Histogram of the distance to the nearest pool at spraint

...

and non-spraint sites 69 Histogram of the amount of overhanging and instream

...

vegetation at spraint and non-spraint sites 70 Histogram of channel margin type and the presence or

...

absence of otter spraints 71 Plot of 14 variables showing three main factors identified by Factor Analysis

...

72

Photos A to F of vegetation physiognomic classes

...

1 13 Box and whisker plot showing percent of forested area

...

in the sub-drainage basins 116 Box and whisker plot showing the mean width of the natural vegetation strip along the five rivers

...

117

Box and whisker plot of the percent of agricultural area within the 50m wide riparian strip in reaches at the three fluvial network zones

...

118

Scatter plot showing the correlation between percent of forested area (BASIN F%) in the drainage basin and channel wetted width

...

119

Scatter plot showing the correlation between proportion of forested area within the 50m wide riparian strip and channel wetted width

...

120

Histogram showing vegetation classes at sites with and without otter spraints

...

121

Large woody debris in the Guarga River

...

144

Woody jam in the Veral River

...

144

Eroding bank in the Ara River ... 145

...

...

Figure 4

.

5

.

Fallen blocks from canyon walls in the Alcanadre River 146

Figure 4

.

6 . Surfacing bedrock in the Guarga River ... 147

Figure 4

.

7

.

Mature patch of white poplars in the Alcandre River

...

147

Figure 4.8. European otter (Lufra Iufra)

...

148

Figure 4.9. A sleepy otter in Pont de Suert rehabilitation centre

...

148

...

Figure 4.1 0

.

General view of the Alcanadre River canyon 149 Figure 4.1 1

.

Turbid water in the lsabena River

...

149

...

xii

Acknowledgments

I would like to thank my multiple supervisors Drs. Richard Hebda, Don Eastman, and Asit Mazumder for their commitment in providing me with direction and

support. In particular, I would like to thank Dr. Asit Mazumder and Dr. Richard

Hebda for their financial support. It would not have been possible to complete this thesis otherwise.

I owe my deepest gratitude to my friend Dr. Rene Alfaro, for his help with statistics and for his continuous support at low times; and to my friend David Badke for his generous help with all my computer troubles, text formatting, and imagery

design; even though he says that he is keeping a tab at 25 cents per

consultation, I am sure I will never be able to pay him back. Also to my friend Sheri, for her committed help in proofing my thesis drafts in between her multiple occupations. As well, I would like to thank numerous people and organizations in Spain, who provided me with airphotos, stereoscope, and readily answered my requests for information or advice, in person and later by e-mail: Ramon Jato, Paloma Barrrachina del Val, Angel Jarne, and Francisco Izquierdo, all from the Departamento de Medio Ambiente, Gobierno de Aragon; Ana Oliva, lnstituto de Estudios Altoaragoneses; Dr. Cesar Pedrocchi, lnstituto Pirenaico de Ecologia;

and Rogelio Galvan, Confederacion Hidrografica del Ebro. I also want to thank

Thomas Gore and Heather Down, from the Biology Advanced Imaging Lab, for

their help with Powerpoint presentations, and to my grad student colleagues

John, Rebecca, and Jill for their help and advice.

I am also grateful to my friends Jim Brander, Thomas Heyd , Bill Eisenhauer, Gabi

Hirt, Jutta Gutberlet, Isabel Leal, and Thomas and Pilar Munson for much help on many issues, and for cheering me up when I needed it; and to my dearest sons Rayn and Nathan for being so patient with their student mom and for finding time to discuss things at length when the need arose.

xiii

Dedication

A

mi

padre

Saltando entre las rocas cantarina,

descansando en 10s remansos, se va volviendo tranquila,

constante, hasta llegar al mar.

Chapter 1

Chapter 1

1 . The issue

The negative effects of watershed and floodplain development on river systems are well known, including extensive changes in physical and chemical

characteristics such as hydrologic regime, channel morphology, and water quality (Bravard and Petts 1996). These alterations usually result in substantial modification of fluvial habitats and of the biological communities dependent on them (Bunn and Arthington 2002). Numerous jurisdictions are concerned in the health of riverine systems and much effort and many resources are being invested into assessing the ecological integrity of freshwater courses and

restoring riverine systems worldwide (Directive 2000/60/EC; Stromberg 2001 ; Lai et

al. 2004; Golumbia et al. 2004).

In the Iberian Peninsula, rivers have been modified by watershed deforestation,

conversion of the land to agriculture (Garcia Ruiz et al. 2001 ), and fragmentation

of riparian vegetation communities for millennia (Costa Tenorio et al. 1998).

These historical changes to river systems may still be having a role today in fluvial form and processes such as channel pattern and flow and sediment regimes

(Gregory and Gurnell 1988; Brown 2002). In more recent times, rivers have been altered by construction of flood defence works, dams, and roads, and by human uses such as mining gravel from the river bed. These alterations impact not only specific sites, but affect large sections of the river upstream and downstream

(Rivier and Seguier 1985). Other changes may be taking place as a result of the rural population decline over the last fifty years. In the Spanish central Pyrenees many people have emigrated to urban centres, and abandoned fields and pastures are being colonized by shrubs and young trees. River systems with flows not regulated by dams, may have quasi-natural hydrology regimes, and their high flows may inundate the adjacent lands unless prevented by defence works.

Riparian vegetation is known to be an important element in maintaining the

stability of fluvial systems (Swanson et a/. 1982; Gregory et a/. 1991 ; Beschta 1991

Tabacchi et al. 2000). The roles of streamside and riparian buffer vegetation are

Chapter 1

providing nutrients to aquatic organisms (Cummins 1975; Cummins et al. 1989;

Thoms 2003), to supplying essential morphological features (such as large organic debris) (Bravard and Petts 1996; Pikgay and Gurnell 1997), and maintaining the integrity of the stream banks (Thorne 1990; Millar 2000). Stream invertebrates, fish communities, riverine mammals and birds have been linked to the condition of

the riparian vegetation (Karr and Schlosser 1978; Cummins et a/. 1989; Kinley and

Newhouse 1997; Robinson et al. 2002).

Although the importance of the riparian zone is recognized, critical measurable attributes of a "good" riparian zone and their relationships to the stream are not well established, especially in regions with a long history of human use such a s

the Iberian Peninsula. This point raises the following questions:

a) Do various structures and widths of riparian vegetation differ in their ability to maintain the biophysical integrity of the study rivers? or in other words, is there a relationship between stream attributes and riparian vegetation characteristics in the region?

b) Can the ecological condition of the riverine habitat be tracked using an relatively easy measured indicator species like the river otter, and

because the otter is in itself of interest and concern? or in other words, can

the otter Lutra lutra L. be used as an indicator of the biophysical condition

of riverine systems in Aragon, Spain? Therefore the purposes of this thesis are:

1 ) to examine whether or not the riparian vegetation is contributing to the

integrity of the aquatic fluvial habitat in Pyrenean and pre-Pyrenean rivers. Specifically, the role of the riparian vegetation as reflected in the morphological attributes of these rivers.

2) To examine if the ecologically symbolic otter be used as a proxy indicator of riverine ecosystem integrity.

2. The riverine ecosystem

In the overall landscape, rivers are thin strands of water. The water from the seas rises as vapour and forms clouds that travel across oceans and fall as rain or snow on the land of the continents. This water is what sustains the life of all terrestrial plants, animals, and humans alike. Once it reaches the ground it flows down the slope, either as subsurface or overland flow, converging and carrying sediment, organic matter, nutrients, and contaminants into rivulets, creeks, streams, and rivers. These thin strands of water form a shimmering network across the landscape. The water strands and their margins, the sole home to an array of plants and animals, constitute the earth's riverine ecosystems.

The water that falls on the land follows millions of paths to converge into rivers. The portion of the landscape from which a river draws its water is its watershed. The watershed therefore has a major influence on the riverine ecosystem. The riparian-fluvial network, on the other hand, acts as the connecting link between

the uplands and the lowlands of the watershed. The river is connected three-

dimensionally to its watershed: longitudinally, laterally and downwards.

The longitudinal connectivity of the river is the most obvious one: water, solutes and sediment collected in the highland's rivulets travel downstream to the wide channel and floodplains of the lowlands. On the other hand, migratory fish, birds and mammals travel upstream, or downstream, at certain times of the year, for reproduction or to complete their development, as part of their life cycle. In addition, the return of anadromous fish to upland streams from the ocean constitutes an important flux of biomass to the riparian-fluvial ecosystem.

Downwards the river is connected to the watershed through the hyporheic zone, or zone of exchange between surface water (flow) and groundwater. The most accepted definition of the hyporheic zone is that of saturated interstitial areas beneath and beside a riverbed and containing some proportion of surface

water from the channel (Edwards 1998). Triska et al. ( 1 989) set the proportion of

surface water at 10% whereas other researchers prefer to define it simply as the area occupied by hyporheic fauna (Standford and Ward 1988). In reaches over extensive alluvial aquifers, this zone can extend laterally from the river channel

and under the floodplain for hundreds of meters (Edwards 1998). The sediment interstices are inhabited by an abundant community of micro-organisms (epilithic biofilm), which are thought to have an important role in organic matter

decomposition and oxygen production for the entire fluvial ecosystem. Therefore this zone is considered an important element for the preservation of surface water quality. In addition, the epilithic layers are an important food source for

invertebrates. The invertebrate hyporheic community is composed of a large

number of species, some of which are subsurface obligates and others are facultative. Because of the many unique traits of species living in the sediment

interstices, the hyporheic zone is considered a hotspot of biological diversity

(Edwards 1 998).

Finally, the river is connected laterally to the uplands through the riparian zone. According to the Webster's Dictionary ( 1 985), 'riparian' denotes 'living or located on the bank of a watercourse, or a lake'. Therefore riparian zones are those areas of land adjacent to streams and other water bodies. According to

Gregory et al. (1991), these areas are ecotones or transition zones between the

terrestrial and the aquatic environments. Often typical flood-tolerant vegetation species occupy these areas. However, not all riparian zones support distinctive vegetation, since in constrained channels the hillslope communities may reach almost to the water's edge, except for a thin band of hydrophytic vegetation. On the other hand, in moist temperate climates the high moisture content in the environment may obscure the distinction between streamside and upland vegetation. Riparian zones have been defined according to various criteria such as presence of hydric soil, or hydrophilic plant associations (Beschta 1991).

Gregory et al. (1991) suggest that definitions based on these criteria alone

encourage a rigid and inappropriate delineation of the riparian zone. Following

Meehan et al. (1 977) and Swanson et a/. ( 1 982), they propose an ecological

definition based on the linkages between aquatic and terrestrial ecosystems, by which 'the riparian zones extend outward to the limit of flooding and upward into

the canopy of the streamside vegetation' (Gregory et al. 1991, p. 540). Along the

same lines, Ward et al. (2002) describe 'riparian corridors' as 'linear features of

Chuptcr

1

6

sea' (Ward et a/. 2002, p.518). No precise boundaries are given for these corridors

except for a general concept of 'narrow canyon-constrained' and wider 'alluvial floodplains' corridors, like 'beads on a string'.

The riparian vegetation is known to contribute to bank stability (Thorne 1990; Abernethy and Rutherfurd 1998), deposition of sediment, infiltration of runoff, and uptake of excess nutrients, often from agricultural lands (Karr and Schlosser 1978; Tabachi eta/. 2000); it is also the main source of allochtonous material to the aquatic system, in the form of coarse woody debris (Heede 1985; Bilby and Bisson

1998), organisms, organic matter (Cummins ef al. 1989; Bisson and Bilby 1998) and

dissolved carbon (Newbold 1992; Thoms 2003); and provides shade, for control of water temperature, and cover for fish from predators (Beschta 1991 ). In addition, the alluvial aquifer beneath floodplains helps maintain the base flow at times of drought. From this account, it becomes evident the numerous linkages between the aquatic and the terrestrial environment assigned to the 'riparian zone' or the 'riparian corridor'.

It is well accepted that riparian vegetation has a significant influence on channel

morphology (Keller and Swanson 1 979; Gregory and Gurnell 1 988; Thorne 1 990;

Abernathy and Rutherfurd 1998; Gran and Paola 2001), and that the roots of streamside vegetation provide additional strength to the banks against the

erosive forces of the flow. A theoretical model developed by Millar (2000) in

which riparian vegetation was a factor, was tested with field data from 137 rivers and was successful in distinguishing meandering from braided channels. The model showed that bank vegetation exerts a quantifiable control on alluvial channel patterns. Also, in an experiment, Gran and Paola (2001) showed that riparian vegetation alone could substantially alter channel geometry. However, in natural rivers, the assignment of cause and effect of channel changes to riparian vegetation modifications has been obscured by simultaneous land-use modifications on the watershed. A fourth-order stream in British Columbia

presented an interesting case. The channel changed from a meandering 30m wide channel to a 150m wide braided channel soon after logging the riparian forest, while there were no noticeable changes to the watershed, which was

mostly a pristine forest within a park (MacVicar 1999). In a study of twenty-six

reaches in urban and rural watersheds, Hession et al. (2003) concluded that

'riparian vegetation exerts a strong influence on channel width regardless of the

level of urbanization in the watershed' (Hession et al. 2003, p.150).

The riparian zone acts as energy sink for the erosive power of the flow and as structural support for the banks. The ground and vegetation cover provide a rough surface area on which to spend part of the flow's kinetic energy, thus reducing its speed and its erosive power. In addition, the root systems of the riparian vegetation facilitate percolation and function as a structure that anchors the substrate, maintaining the stability of the banks (Bannerman 1998; Bechsta 1991). In this process, the density of vegetation and ground cover plays an important role in the rate of water percolation and in preventing particle entrainment (Abernethy and Rutherford 1998). Various studies indicate that a thick herbaceous ground cover is most efficient for preventing surface erosion and promoting percolation whereas a dense canopy layer is most effective for providing bank structural support and maintaining the stability of the banks (Lewis and Kovacic 1 993).

Another important role of vegetation in fluvial systems concerns the supply of organic matter and dissolved organic carbon to the aquatic ecosystem. Fluvial systems are largely dependent on an allochtonous supply of organic matter for their energy (Jones 1975; Suberkropp 1998). Jones ( 1 975) quotes studies by Mann

( 1 969)and by Westlake et al. ( 1 972) which showed that in two British rivers,

allochtonous inputs were responsible for 50% and 88% of total inputs. The

decomposition of leaves, twigs, and other organic matter by fungi and bacteria, makes detritus available to higher trophic level organisms such as invertebrate detritivores. Dissolved organic carbon, the other main component of the aquatic food web, originates from leachates and exudates of terrestrial and aquatic organisms, from soil carbon of the floodplain or riparian zone, and from groundwater entering the stream through the hyporheic zone (Suberkropp 1998).

Dissolved organic carbon is

the

dominant form of organic matter in river water;

its

Chapter 1

invertebrates (detritivores) (Newbold 1992). Recent research shows that regular flooding of floodplains is the main source of dissolved organic carbon (Thoms

In addition, forested riparian zones have an influence on channel morphology and aquatic habitat complexity by supplying structural elements in the form of

large woody debris. Many studies have shown that large woody debris is an

important element of channel morphology (Sedell and Froggatt 1984; Triska 1984;

Heede 1985; Piegay and Gurnell 1997; Abbe and Montgomery 1996; Edwards et

al. 1999; Piegay et al. 2000). However, the morphological role of large woody debris (LWD) depends on its type, size and position, as well as on the size of the channel. Logs with the root wads attached have been found to be more able to withstand flow than those without. Riparian trees that fall into the channel due to windthrow or because of riverbank erosion constitute the main source of logs with root wads (Abbe and Montgomery 1996). Large logs and woody debris accumulations within the channel dissipate kinetic energy, enhance sediment and organic matter deposition, create pools, provide fish cover and habitat

substrate and nutrients for aquatic organisms (Bisson et al. 1981 ; Bilby 1984).

Woody debris accumulations on top of gravel bars have been found to be associated with the development of vegetated islands and forested floodplains

(Abbe and Montgomery 1996; Edwards et al. 1999; Gurnell and Petts 2002),

whereas logs along the banks buttress the banks against erosion and create backwater ponds and areas of sediment deposition (Keller and Swanson 1979). Abundance of LWD has been associated with increased aquatic habitat

diversity and fish production (Angermeier and Karr 1984; Bisson et al. 1987; Flebbe

1999). LWD also offers shelter to fish from predators (Allouche 2002). Fish within pools associated with LWD tended to remain in the pools and move less than fish

in pools without LWD (Harvey et al. 1999).

The riparian zone acts as a 'nutrient filter and sink', with riparian plants taking up and storing excess nutrients from soil and groundwater. The symbiotic association of plants with microbes, such as mycorrhizal fungi, enhance the nutrient

Chupter

1

and soil. The increased amount of organic matter together with moist soil

conditions can potentially intensify the activity of denitrifying bacteria (Hill 1996;

Tabachi et al. 2000; Groffman and Crawford 2003). Cole ( 1 981 ) has suggested

that nitrogen uptake is often limited by the availability of the resource. For

example, black poplar trees (Populus nigra) when fertilized can take up as much

as 400 kg of N per hectare per year, whereas in the natural environment they

only assimilate 16 kg per hectare per year (Cole 1981 ). The soil type, depth to the

groundwater, surface area, and type and density of the vegetation cover are important elements affecting the efficiency of the riparian zone for nutrient

uptake. A review by Hill (1 996) of nitrogen uptake under various hydrologic

conditions and riparian vegetation communities shows that with a shallow aquifer under the riparian zone, a riparian band of deciduous forests 20m wide can remove 90% of nitrates, whereas a band 50m wide or more can remove from 98 to 100%. However, riparian forests over aquifers deeper than 4m are not as efficient. Riparian organic soils, on the other hand, with an abundant supply of carbon, have a higher nitrate removal rate than more mineral soils, related to

higher activity by denitrifying bacteria (Hill 1996; Groffman and Crawford 2003).

Productivity of fish is determined by abundance of habitat and food resources. Both fish habitat and fish prey resources are associated with the structure and

composition of the riparian zone (Cummins et al. 1989; Gregory et al. 1991). In

addition, backwater channels and ponds in the floodplain constitute an

important habitat for riparian and aquatic species (Muhar and Jungwirth 1998),

and the associated alluvial aquifer is an important element in maintaining a base

flow during the dry season.

So far, I have reviewed the various mechanisms by which the aquatic ecosystem depends on materials, structure and processes of the riparian zone. However, riparian communities also depend on materials and processes of the aquatic habitat. Riparian plants depend on accessible groundwater and/or periodic flooding for their survival and reproduction, and fish and aquatic insects are the main food prey for many terrestrial and riparian species. Riparian consumers such as birds, bats, spiders, amphibians, reptiles and mammals benefit from energy

transfers from the aquatic ecosystem in the form of aquatic insect emergence as well as fish and macro-invertebrates. In a temperate forested stream Nakano and Murakami (2001) showed that aquatic insects provided a significant proportion of the food intake to riparian insectivorous forest birds during the defoliation period, and 25.6% of the annual total energy demand. Similarly terrestrial insects contributed greatly to the energy demand of fish species, a

total of 44% of the annual energy budget; the dependence of fish on terrestrial

prey being greatest during the summer. These authors suggest that the reciprocal prey flux across the two habitats, and the asynchrony of the maximum flux

(summer from forest to stream, and defoliation period from stream to forest), provides a reciprocal subsidy of prey to predators in both habitats at times of scarcity. This study further exemplifies the ecological importance of the linkage between the aquatic and riparian habitats.

The riparian zone is thus intrinsically connected to the river channel, and plays a fundamental role in the functioning of the hydrological, geomorphological, chemical and biological processes of the fluvial system, as an extensive body of research shows (Vannote et al. 1980; Swanson et al. 1982; Gregory and Gurnell

1988; Junk et al. 1989; Decamps 1996; Tabachi et al. 2000; Gomi et al. 2002; Reeves et al. 2003; May and Gresswell2003; Thoms 2003). In addition, many studies show that riparian ecosystems have greater species diversity and

population abundance than adjacent uplands (Gregory et al. 1991; Naiman et

a/.

1993). Riparian ecosystems are home to numerous species of plants,

invertebrates, reptiles, amphibians, birds, and mammals, many of which,

particularly in arid and semi-arid regions, are riparian obligate species, requiring the special conditions of these areas for their survival (Naiman et a/. 1993; Kelsey and West 1998; Pollock 1998). Yet, it is estimated that riparian ecosystems amount

to 1 to 3 percent of the terrestrial landscape (Patten 1998). The rarity of these

ecosystems furthers the cause for their protection and restoration.

3. Effects of watershed development on riverine systems

When the water held in the atmosphere falls as rain or snow, a portion is intercepted by vegetation and eventually evaporates. Other portion of

precipitation reaches the ground directly, or after trickling down stems and

leaves. A fraction of the water that reaches the ground will be used b y the plants

and will again reach the atmosphere in the form of vapour through the physical process of evapo-transpiration. Clearly, the denser and more extensive the cover of vegetation over the landscape, the less water will reach the fluvial network. In addition, plants use part of the water in the soil during the transpiration process.

A portion of the precipitation that reaches the ground percolates into the soil.

This water moves downward, and becomes part of the groundwater, a most

important fresh water reservoir (Beaumont 1975), or laterally downslope as

subsurface flow, when a less permeable layer is encountered. When

precipitation exceeds the infiltration capacity, or as the soil becomes saturated to the surface, the water will flow downslope as overland flow, commonly called runoff. In well-developed soils, with high organic matter content and a dense vegetation cover, precipitation may never exceed the infiltration capacity; high overland flows being associated with semi-arid areas, thin soils and sparse

vegetation (Knighton 1 998).

Saturated soil has a low cohesion and it is easily moved by water either in suspension or creep modes. Water soluble substances in the soil will also be carried in solution with the runoff. The greater the volume of runoff, the steeper the slope, and the greater the erodibility of the material, the higher the amount of sediment that will move downslope on its way to creeks and rivers.

Rivers, at the lowest profile of the drainage basin, concentrate the water, organic matter, sediment and dissolved substances carried downslope by the movement of water. Vegetation cover is a major control over the amount of sediment and runoff reaching the rivers. Deforestation accelerates erosion on slopes, increasing the rate of gully formation, which together result in higher rates of sediment and runoff being supplied to rivers. Conversion of forests to

agricultural land has been associated with 3 to 5 times higher recurrence of

floods (Knox 1977). In the United Kingdom, the suspended sediment

concentrations in rivers coursing near construction works have been noted to be 2 to 10 times, and sometimes up to 100 times, greater than under normal

conditions (Walling and Gregory 1970). The higher volume of runoff and sediment supplied to rivers generally causes aggradation of the channel bed, resulting in up to 300 percent wider (Brooks and Brierley 1997), shallower and less sinuous

channels (Starkel 1991

:

Knighton 1998; MacVicar 1999). Deforestation, and more

recently urbanization (or more properly 'asphaltization'), has caused severe changes on rivers throughout the world (Beaumont and Atkinson 1969; Starkel

1991 ; Brooks and Brierley 1997; MacVicar 1999). It is evident, that the physical and

chemical characteristics of the riverine system, comprising the river channel, and its associated riparian zone, floodplain and wetlands, as well as the plant and animal communities that depend on it, are directly linked to processes and

human activities taking place on its drainage basin (Richards et a/. 1996). The

deposition of large quantities of sediment in parts of river channels may change entirely the nature of the channel; the turbid water may disrupt the animal and plant life of the fluvial ecosystem; and toxic substances present in sediments may bioaccumulate in tissues of organisms from algae and microorganisms to higher trophic species of the food chain, causing growth and reproductive problems, declines of populations, and local extirpation of species. The river is thereby linked to and highly influenced by its watershed and floodplain.

4. The influence of vegetation on channel morphology

Riparian communities vary greatly across geographic regions when species composition is considered (Patten 1998). Although there is consensus concerning the important role of vegetation on the functioning of fluvial ecosystems, there are different opinions about the most efficient vegetation community for specific functions such as supporting banks and maintaining deep and narrow channels. These functions are of interest because channels with those characteristics are potentially better for fish habitat than wide and shallow ones. There is a general view that trees afford the banks the most protection against the erosive power of the flow, and numerous regulations and restoration projects are aimed at

maintaining or establishing trees along riverbanks (Lee et al. 2004; Schaff and

McLeod 2004). There is controversy however, about the vegetation type that provides the best riverbank support. Various studies have shown that grassy

Chapter- 1

banks result in narrower and deeper channels than forested banks (Murgatroyd and Ternan 1983; Trimble 1997; Lyons 2000). Other studies show that forested

banks result in narrower and more stable channels (Charlton et a1 1978; Stott

1997; Hey and Thorne 1986). Liquori and Jackson (2001 ) working on streams of the

east side of the Washington Cascades, found that reaches in shrublands had narrower and deeper channels with more undercut banks than forested reaches. As several authors point out, a major factor responsible for stream bank support is

the proportion of root mass to soil mass (Smith 1975; Liquori and Jackson 2001 ;

Micheli and Kirchner 2002). When considering the type of vegetation that will most effectively protect the banks from erosion, factors such as the physiognomy of particular species involved and local environmental and physical

characteristics need to be considered. These factors include stem type and density, bank height relative to the roots depth, bank composition, the

environmental setting (wet, dry, warm, cold, or prone to frost) and river size (Smith 1975; Thorne 1990; Montgomery 1997; Stott 1997; Abernethy and Rutherford 1998). For example, in frost affected areas forests will moderate the temperature

and the soil will be less prone to frost than if covered with grass (Stott 1997),

whereas in regularly saturated soils tree species not adapted to those conditions will tend to develop roots that remain above the groundwater table and will be shallow-rooted and prone to windthrow (Grinel and Wolff 1998). Generally, in non-cohesive banks, trees facilitate soil drainage and support the banks better than grass (Thorne 1982, 1990).

Forested riparian corridors are also a source of large woody debris, and the effect of large logs and debris jams on bank erosion and channel width is also controversial. Some authors have pointed out that presence of LWD in the

channel increases erosion (Trimble 1997; Liquori and Jackson 2001 ) because of

the turbulence generated around the obstruction (Keller and Swanson 1979), resulting in wider channels; and also uprooted trees create a site of erosion on the banks. On the other hand, many authors indicate the effect of LWD in protecting the banks by enhancing sediment deposition and buttressing the banks (Keller and Swanson 1979; Gregory and Gurnell 1988; Grinel and Wolff

Chapter

1

1998; Naiman et

a/.

1998; Bilby and Bisson 1998). Grinel and Wolff ( 1 998)

measured the amount of sediment released to the stream by uprooted trees in forested buffers heavely affected by winthrow ( 33% of the trees). They also measured the volume of the sediment wedges in the channel, 93% of which was initiated by LWD. At half of the sites the sediment stored in wedges was ten times that of the erosion caused by the uprooted trees. Montgomery ( 1 997) points out that forested channels tend to have a higher diversity of channel form than channels through grasslands. Thus it would seem that bank erosion and increased channel width caused by LWD tends to be local.

Though under certain environmental or soil conditions shrubs or grasses may afford as much or more protection to the banks than trees. The Alexandra River, in Banff National Park, an anastomosed channel with a bankfull discharge of 85m3/sec, with cohesive banks of 70% silt, and 30% clay and fine sand, has been without noticeable channel migration for the last 2,500 years (Smith 1975). Smith

( 1 975) attributed the stability of the channel to the density of roots ( 1 6 to 18 % by

volume of soil), which due to the low rate of bacterial decomposition had accumulated in the soil to a depth of 7.6m, and to the 5cm of root mat on the surface. The accumulation of grass and willow roots in the soil and 5cm root mat on the surface provided 20,000 times more resistance to erosion than the same banks without the vegetation (Smith 1975). Other authors also indicate that grass covered surfaces facilitate greater rates of sediment deposition than forest or

shrublands (Trimble 1997; Allmendinger et a1 1999). However, in warmer climates

channel migration rates tend to be higher in grasslands compared with forested

lands (Smith 1975; Allmendinger et al. 1999). Also riverbanks colonized by hydric

meadow vegetation (sedges- Carex sp. and rushes- Juncus sp.) have a higher

root density per soil mass (50%) and are five times more resistant to erosion than banks colonized by xeric communities of annual grasses and scrub, with a 5% root biomass to soil mass ratio (Micheli and Kirchner 2002).

The strength afforded to the banks by vegetation has significant implications on channel morphology, contributing to channel cross-section form and channel

According to Church (2002) multiple-thread wandering gravel channels are an intermediate state between single-thread meandering and multiple-thread braided channels. Wandering channels have irregularly sinuous individual branches separated by stable vegetated islands. An analytical

hydrogeomorphic model developed by Millar (2000), which accounts for the influence of bank vegetation, successfully identified braided and meandering

patterns in a set of 137 natural rivers. An experiment by Gran and Paola (2001 )

about the effect of vegetation on channel morphology demonstrated the consistent influence of vegetation density on channel geometry and form. As vegetation density increased, bank stability increased, lateral mobility, number of branches, scour hole mobility and width to depth ratio decreased, and scour hole depth increased. Under the same conditions of slope and discharge, with the lowest vegetation density treatment the model in plan view resembled a braided channel, whereas under the highest vegetation density it resembled a wandering river, with a few individual branches separated by large vegetated

islands. A study by Tooth and Nanson (2004) explained the divergence of

channel pattern predicted on hydraulic and sedimentary discriminating diagrams in two rivers in central Australia, from meandering and braided to a straight channel, as a result of the stability of the banks provided by the vegetation (Tooth and Nanson 2004).

5.

Use of indicator species for ecosystem evaluation

Indicator species have been used for decades as a means to evaluate

environmental conditions (Thomas 1972). For example, plants and invertebrates have been used to assess air and water quality (Philips 1980). In general, an

indicator species is an organism whose presence and population fluctuations

can be used as an index of the environmental conditions of the habitat,

ecosystem or community of interest (Meffe and Carroll 1997; Zacharias and Roff 2001). However, the same general concept has been referred to using the terms 'umbrella', 'charismatic', 'flagship', 'keystone', 'sentinel', and 'focal' species (Meffe and Carroll 1997; Simberlloff 1998; Zacharias and Roff 2001). Zacharias and Roff (2001) suggest that the term 'focal' species includes all the other types.

Chapter-

1

16

They define a 'focal species' simply as a valuable species, either for ecological or social reasons, or for conservation or management purposes; and 'sentinel' species as being synonymous of 'focal' species. The other terms are defined in

Table 1 .

Table 1 .I. The four categories of focal species and their definitions (Zacharias and Roff 2001).

Concept Definition

Keystone A species that has a considerable influence in the community, far beyond what would be expected given its biomass or abundance, and whose removal would lead to further loss of species in the

community (Simberlloff 1998; Zacharias and Roff 2001 ).

Umbrella A species with large area requirements, and its presence in the environment would indicate that other species with lesser area requirements would also be present. Therefore the protection of the habitat of an umbrella species would protect other species as well.

Flagship A charismatic species that arouses public support for its conservation, which generally entails conservation of its habitat.

Indicator A species whose presence can be used to identify a particular habitat, community or ecosystem; or the habitat quality for other species, communities or ecosystems.

The indicator species concept is the most vague. The US Forest Service has been

a main promoter of its use since, by law, each National Forest in the US must

identify 'Management Indicator Species (MIS)' (Code of Federal Regulations

1985; quoted by Landres et

al.

1988). Species selected as MISS include: 1 ) rare,

threatened and endangered; 2) of social or economic value; 3) with habitat

requirements especially sensitive to human activities; 4) ecological indicators (i.e., species used to monitor environmental, habitat or community conditions). As can be appreciated from this list, there are various roles and definitions assigned to indicator species. The lack of consensus about the scientific theory behind the concept, and on the definition, selection criteria, and application

standards of indicator species has sparked considerable debate (Landres et al.

1988; Simberloff 1998; Zacharias and Roff 2001).

The main criticism is well exemplified by Simberloff ( 1 998):

'the concept is problematic because there is no consensus on what the

the best indicator species even when we agree on what it should indicate'

(Simberloff 1998 p. 247).

Evidently, a major motivation for the use of indicator species is its simplicity, in contrast to having to evaluate a series of abiotic and biotic attributes of the ecosystem of interest. In spite of the debate surrounding the use of indicator species, all of the above authors consider the application of the indicator species concept useful for conservation and management.

Landres et a/. ( 1 988) recommend that a species selected as an ecological indicator has the following characteristics: sensitive to changes in habitat

attributes; exhibits low levels of variability in relation to the factors of interest; and

is a permanent resident. Zacharias and Roff (2001 ) separate indicator species

into two categories; composition indicators and condition indicators. A composition indicator is one whose presence, absence or abundance is a

surrogate for a particular habitat, ecosystem or community; whereas a condition indicator indicates the quality of the habitat, ecosystem or community.

Condition indicators are those used to monitor environmental change, whereas composition indicators are used to identify representative areas. According to Zacharias and Roff (2001), part of the confusion about definition and application of indicator species seems to stem from having these two types of indicators under one category. An important characteristic of both types is that an indicator species exhibit a low temporal and spatial variability and be independent of spatial scales. Along the same line, Meffe and Carroll (1 997) suggest that in selecting condition indicators, long-lived species and those whose populations fluctuate greatly should not be selected, since they may provide misleading signals. In addition, a condition indicator is expected to be able to differentiate between natural and anthropogenic disturbance, and to be able to provide an assessment over a range of stress. On the other hand, a composition indicator is expected to exhibit a specific range of ecological tolerances and consistently be part of a specific community or habitat type (Zacharias and Roff 2001).

An interesting approach to the use of focal species for conservation

management was proposed by Lambeck ( 1 996). The author proposes a multi-

species umbrella system by which all species considered at risk in an area are identified and each allocated to at least one of four risk categories: area- limited, resource-limited, dispersal-limited and process-limited. Within each category the most sensitive species or the one with the most demanding

requirements is used to define the acceptable level at which that threat can be

allowed to occur or the minimum value for meeting those requirements. By using the most demanding species in each category to set the minimum parameters, the landscape will encompass the requirements of all the other species.

In the Iberian Peninsula, the Eurasian otter Lutra lutra L. is a top-of-the-food chain species in the river ecosystem, using the riparian zone for refuge, resting and reproduction. According to various authors of otter studies, the European otter is a good indicator of the condition of fluvial-wetland ecosystems because of its

sensitivity to pollution and habitat alteration (Mason and Macdonald 1986; Ruiz-

Olmo and Delibes 1998). The otter has been legally protected in Spain since

October 1973 (Decree 2573173) and deliberate killing of otters by humans

relative to total otter mortality decreased from 90% in the years 1950 to 1980 to

16% in the years 1990 to 1996 (Ruiz-Olmo et al. 1998). This protection measure has

likely contributed to the increase of the otter population in Spain in the last two decades. One of the areas of expansion has been the province of Huesca,

where the area occupied by the otter tripled from 1984 to 1996 (Ruiz-Olmo et al.

1 998).

The otter's diet is composed mainly of fish (Callejo and Delibes 1987; Ruiz-Olmo et

al. 2001), and in healthy environments, otter populations are limited by fish prey resources, food availability affecting both otter mortality and reproduction (Kruuk

1995). Lannon and Reynolds ( 1 991 ) have suggested that river otters may serve as indicators of rivers' ecological condition for its sensitivity to contaminants and alteration of their habitat. The otter meets various characteristics of a condition

indicator species (Meffe and Carrol 1997): it is a resident, specialized, and short-

Chapter-

1 19 fluctuations. These charcteristics may require clarification. Otters are permanent

residents of those areas they inhabit, and while they can live up to 15 years in

captivity, in the wild their life expectancy is approximately three years (Kruuk 1995). Therefore otters may be considered short-lived rather than long-lived animals. Also, even though the otter may prey on other species other than fish, a

study by Kruuk ( 1 986) showed that the otter is one of the most specialized

carnivores.

6. Hypotheses

According to the literature, both living and non-living vegetation from the riparian zone can have a significant effect in channel morphology of alluvial rivers. Therefore, I expect that characteristics of the riparian zone will be reflected on biophysical attributes of the Pyrenean and pre-Pyrenean rivers. Specifically, alluvial reaches with a higher proportion of trees and shrubs in the riparian zone will be expected to have more stable banks, thus narrower and deeper channels (or width to depth ratio), and higher in-stream vegetation cover than reaches with a lower proportion of these vegetation types. In addition, a higher proportion of mature trees in the riparian zone will contribute to reaches with larger amounts of large woody debris, which in turn will increase the amount of pool habitat. However, because of the known influence of the watershed on the biophysical attributes of river systems, watershed

characteristics will also need to be considered.

Morphological and structural features of a river channel are key elements of the aquatic habitat, contributing to the requirements of the various life stages of many aquatic organisms. Since the otter relies primarily on fish, the existence of a relatively abundant otter population indicates an abundance of fish, which in turn, reflects the condition of the aquatic habitat, and because the otter is also dependent on the vegetation cover of the riparian zone for resting and

reproduction, it can be considered an indicator of the good overall condition of the riverine system.

Chapter

1

20

1 ) Do various types of structure and width of the riparian vegetation differ in

their ability to maintain the biophysical integrity of fluvial systems, within the constraints

systems?

2 ) Can the otter I

that watershed characteristics may impose on river

.&a lutra L. be used as an indicator of the biophysical condition of riverine habitats in Aragon, Spain?

7 . Approach

To answer the research questions, I measured characteristics of the riparian

vegetation, channel morphological attributes, and conducted an otter survey in twenty-two reaches of five rivers in Huesca, Spain. Additionally, I delineated the boundaries and quantified land-use cover of each reach drainage basin. First, I examined the relationship of otter relative abundance to stream morphological

features (Chapter 2); and secondly, I examined the relationship of the riparian

vegetation to stream attributes (Chapter 3). Because riparian vegetation has a

direct connection to otter ecology, I also examined relationships between

characteristics of the riparian buffer and otter abundance (Chapter 3). Potential

impacts of catchments' characteristics on otter populations as well as relationships between catchments' and stream characteristics were also

examined in Chapter 3. Chapters 2 and 3 provide overall details of the study

area, the study's design and methods used. In Chapter 4,l discuss the results of the study and examine future research needs. Additionally, I consider the potential implications of the results for the restoration of rivers in Huesca, Spain.

Chapter

2

The relationship of European otter Lutra lutra

L.

to

morphological characteristics of fluvial systems of

Huesca, Spain.

1. Introduction

Otter populations declined substantially from 1960-80 throughout Central and Western Europe, with the species almost extirpated in Belgium, western Germany, Netherlands, and Switzerland, and in extensive areas of England, Denmark, Sweden, Austria, Italy and France (Mason and MacDonald 1986). In Spain, from

1966-1 985 the population decreased 60% (Blanco 1998). The otter was lost from the Mediterranean coastal drainages, and from large urban, industrial, and intensively cultivated areas, remaining in the western half and in fragmented parts of the original range in central and northeast portions of the country (Delibes and Rodriguez 1990). Several reasons have been suggested for the decline, the most critical being pollution, hunting, and destruction of their habitat (Mason and MacDonald 1986; Ruiz-Olmo and Delibes 1998).

Since the late 198O9s, otter populations have experienced a moderate increase in

much of central and western Europe (Ruiz-Olmo and Delibes 1998; Brzenzinski, et al. 1996; Strachan and Jefferies 1996; Erlinge 1995; Rosoux et. al. 1995). In Spain, otter populations have revived in several regions, according to the results from the last two National Otter Surveys in 1994-96 and 1984-85, In particular, the species showed an expansion in mountainous zones and has recolonized several Pyrenean and pre-Pyrenean drainages (Ruiz-Olmo and Delibes 1998). Although these

expansions are encouraging, the status of otter in Spain is still a matter of concern. The conservation of otter requires a sound understanding of its ecology, especially habitat relationships and food resources, given the importance of these factors in previous population declines. Therefore, the purpose of my study is to examine the relationship of elements of fluvial morphology important for fish habitat to the abundance of European otter in north-eastern Spain. My approach was to examine otter abundance and features of their habitat in a sample of reaches of five intermediate-sized rivers.

The otter is a top-of-the food chain species, and its diet is based primarily on fish (Clavero et al. 2003; Mason and Macdonald 1986; Kruuk 1995; Blanco 1998), and numerous studies show a strong correlation between otter abundance and fish

Granado-Lorencio 1996; Kruuk 1995; Kruuk et al. 1993; Green et al. 1984; Green

and Green 1980). Kruuk ( 1 995) suggests that a stream needs to have a minimum of

8-9 gm of fish biomasslm2 for otters to survive. Therefore, recolonization only takes

place in drainages with adequate fish resources, and those areas with relatively high fish biomass are expected to have a high otter density.

Various characteristics and elements of stream morphology are known to be

important features of fish habitat (Alluche 2002; Reeves et al. 1998; Garcia de

Jal6n et al. 1996; Lonzarich and Quinn 1995; Schlosser 1991 ; Bisson et al. 1987;

Bisson e t a / . 1981; Brown 1975). Among various habitat units identified in river

systems-pools, riffles, rapids, cascades, and glides-the main units are pools and

riffles, the others being subtypes of these two (Church 1992). Fish, particularly large

fish, mainly inhabit pools because of the slower current and the protection

afforded by deep water (Schlosser 1988). Deep pools allow fish species and their

age classes to 'stack' within the water column (Bisson et

a/.

1987). In

Mediterranean rivers, the presence of pools is critical for survival of fish during low

summer flows. Riffles, on the other hand, contribute to the incorporation of dissolved oxygen: highly oxygenated waters are an important requirement, especially for salmonids. Riffles also are an important habitat for benthic invertebrates, an essential food item for many species of fish. Features that increase habitat complexity (roughness) and provide cover for fish such as

instream and overhanging vegetation, large woody debris and boulders have also

been associated with higher fish diversity and biomass (Angermeier and Karr 1984;

Bowlby and Roff 1986; Schiemer and Spindler 1989; Lonzarich and Quinn 1995;

Reeves et al. 1 998; Lehane et al. 2002).

Large roughness elements such as large woody debris and large boulders both provide cover for fish and influence channel morphology. The important role that

large woody debris (LWD) plays in the ecology and morphology of forest streams is

now widely recognized (Keller and Swanson 1979; Bisson et al. 1987; Piegay and

Gurnell 1997; Bilby and Bisson 1998; Naiman and Bilby 1998; Hauer et al. 1999;

Gurnell et al. 2000; Lehane et al. 2002), and numerous studies have shown a

populations (Lehane et al. 2002; Roni and Quinn 2001 ; Rosenfeld et al. 2000; Flebbe and Dolloff 1995). Although most of the studies on LWD have been conducted in North America, particularly in rivers flowing through coniferous forests of the Pacific Northwest, many European studies support the re-introduction of LWD as part of

the efforts to restore European fluvial ecosystems (Lehane et al. 2002; Gurnell and Petts 2002; Ward et al. 2000; Piegay et al. 2000; Edwards et al. 1999; Piegay and Marston 1998; Piegay and Gurnell 1997). For example, in the Douglas River, Ireland, the addition of debris structures resulted in a change in channel morphology with an increase in the amount of pools, eddies and backwaters. The sections with LWD showed a significant increase in trout (Salmo trutta) density and biomass

compared to control sections two years after the installation of the structures (Lehane et al. 2002). European forest trees cannot replicate the influence that the large conifers of the Pacific Northwest have on the streams of that region,

however, experimental studies have shown that even smaller and less decay- resistant hardwood species such as alder can effectively contribute to the improvement of fluvial aquatic habitat (Keim et al. 2002).

Large rocks and boulders are also elements known to contribute to cover for fish (Streubel and Griffith 1993). Large boulders often promote scour and foster the creation of pools. In some rivers, large rocks and boulders create an important array of habitat complexity (Warren and Kraft 2003), with deep pools and crevices for fish cover and invertebrate production.

Other factors being equal (water quality, fishing pressure), the abundance of important features associated with fish habitat quality can be expected to relate to a higher density of fish and, in turn, to a higher density of otters.

2. Study area

The general study area is situated in the province of Huesca in north-eastern Spain. Within this area, I selected sections of five intermediate-sized rivers: the Alcanadre,

Ara, Isabena, Veral and Guarga Rivers (Figure 2.1 ). The river sections range from

30.7 to 1 1.5 km and have similar hydrologic characteristics ranging from 4.1 to 14.3