Universidade de Lisboa Faculdade de Ciências Departamento de … · Universidade de Lisboa Faculdade de Ciências Departamento de Biologia Animal Técnicas e Metodologias de Investigação

Universidade de Lisboa

Faculdade de Ciências

Departamento de Biologia Animal

Técnicas e Metodologias de Investigação II

João Paulo da Silva Medeiros

Relatório

Doutoramento em Ciências do Mar

2012-2015

Universidade de Lisboa

Faculdade de Ciências

Departamento de Biologia Animal

Técnicas e Metodologias de Investigação II

João Paulo da Silva Medeiros

Responsável pela formação:

Maria Teresa Marques Ferreira da Cunha Cardoso

(Professora com Agregação)

Doutoramento em Ciências do Mar

2012-2015

Técnicas e Metodologias de Investigação II – João Paulo Medeiros i

Index

Text

Page

I. Introduction …………………………………………………………………………………………………………… 1

II. Objectives …………………………………………………………………………………………………………….. 2

III. Materials and Methods ………………………………………………………………………………………… 3

3.1. Study areas …………………………………………………………………………………………………… 3

3.2. Methodology

…………………………………………………………………………………………………

3

IV. Results …………………………………………………………………………………………………………………. 4

4.1. Defining Reference Coordinate System …………………………………………………………. 4

4.2. Georeferencing images …………………………………………………………………………………. 4

4.3. Images resampling ………………………………………………………………………………………… 6

4.4. Structure of riparian vegetation and land use …..…………………………………………… 7

V. Supportive literature …………………………………………………………………………………………….. 8

VI. References …………………………………………………………………………………………………………… 9

VII. Acknowledgements …………………………………………………………………………………………….. 9

Figures

Page

Fig. 1. Location of the study areas

……………………………………………………………………………….

4

Fig. 2. Example of the georeferencing process (Touvedo)

…………………………………………..

6

Fig. 3. Comparison of old aerial photos (1960s) and recent satellite images

……………...

7

Projeto OASIS

Técnicas e Metodologias de Investigação II – João Paulo Medeiros Página | 1

I. Introduction

A riparian zone can be described as the narrow strip of transitional land between upland habitats and

perennial or intermittent water bodies. A healthy riparian system supports a great diversity of upland

and wetland-adapted plant species and provides habitat for many different species. Although they

often comprise only a small percentage of total land area, riparian zones represent a vital element in

the overall landscape, acting as both a buffer and an ecological link between water- and land-based

ecosystems. Because riparian corridors represent the area where upland and aquatic habitats merge,

well-developed riparian corridors tend to contain a relatively high degree of diversity, having both

upland- and aquatic-dependent species present. Beyond their high diversity, riparian corridors are

very dynamic and biophysically complex. These habitats encompass sharp environmental gradients,

ecological processes and communities (Naiman et al., 1993). Due to periodic flooding, these corridors

also tend to be rich in nutrients and very diverse in both structural characteristics and vegetation,

thus creating and sustaining a diversity of microhabitats. These microhabitats cover the stream

channel and part of the terrestrial landscape from the high water mark towards the uplands, where

vegetation may be influenced by elevated water tables or flooding and by the ability of soils to hold

water (Naiman et al., 1993). The width of the riparian corridor, the level of control of the streamside

vegetation has on the stream environment and the diversity of functional attributes (e.g. information

flow, biogeochemical cycles) are related to the stream size, the position of the stream within the

drainage network, the hydrologic regime and the local geomorphology (Naiman et al., 1993).

The riparian corridor is frequently disturbed by floods and debris flows, creating a complex shifting

mosaic of landforms over a considerable spatial scale (Salo et al., 1986; Swanson et al., 1988). The

great variety of impacts to riparian areas outlined below often work in combination to further

degrade a particular site or an extended length of the riparian corridor and downstream aquatic

systems (Kattelman and Embury, 1996). Some impacts are chronic or persistent, other effects are

short-term or periodic. The different impacts usually result in a simplified system with fewer species

or individuals or reduced habitat diversity; multiple impacts generally increase the fragmentation of

habitat (Kattelman and Embury, 1996). Stream flow regulation by dams alters channel conditions on

all downstream reaches that may also have local impacts, such as logging or grazing. Construction of

a dam and filling of a reservoir eliminate riparian habitat directly. Downstream of a dam, riparian

conditions are affected by project-induced changes in water availability, substrate and flood

frequency and magnitude. Most water projects alter the natural hydrograph and create a different

regime of seasonal water availability for riparian vegetation. Some common effects of construction

and operation of dams on riparian vegetation include:

Projeto OASIS


� breaking the continuity of riparian habitat and its wildlife migration routes;

� loss of vegetation to roads, penstocks, canals, transmission lines, dams, and other facilities;

� temporary loss of vegetation during construction activities;

� non-native plants may encroach on areas cleared of native vegetation;

� loss of upstream vegetation under continual or seasonal inundation;

� where peak flows are reduced, riparian cover can increase (except on very steep or flat reaches);

� reduced flooding allows formerly flood-suppressed plants (often non-native) to flourish;

� colonization of the formerly active channel can stabilize sediments there and constrict the

channel;

� during very large events, flood levels may be increased by the decline in channel capacity;

� lack of routine floods can reduce seed dispersal and germination of many native plants;

� when discharge during low flow periods is reduced, there is less recharge of floodplain aquifers;

� when less water is available, moisture stress can decrease riparian cover, especially on uphill

edges; in such cases, the riparian area will contract toward the channel;

� more xeric plant species may move into the outer margin of the formerly wet soils;

� when water availability is limited, deep-rooted trees are favored over shallow-rooted species;

� channel incision caused by stream rerouting during construction may lower ground water levels;

� delivery of sediment from upstream is reduced and fine-grained substrate may be lost eventually

(in Kattelmann and Embury, 1996).

To assess flows needed to maintain riparian health, unaltered ecosystems within the same ecoregion

and with the same hydrogeomorphology are needed for comparison. Unfortunately, many rivers

have no such unaltered analog. Furthermore, historical records are very scarce (Alldredge and

Moore, 2012). Without historical vegetation data, however, it is difficult to relate changes in

vegetation composition to hydrology patterns downstream of dams. Old aerial photos are necessary

for applications that involve detection of land use changes over time and/or detection of

constructions or other topographic features.

II. Objectives

The main goal of the present work is to develop and apply remote sensing methods and geographic

information system (GIS) tools to assess the impact of dams in the riparian forests.

Projeto OASIS


III. Material and Methods

3.1. Study areas

Three dams were selected to develop the present work: Fronhas, Touvedo and Vilarinho das Furnas

dams (Fig. 1). One of the factor that weighed in the choice of the dams listed was the availability of

historical aerial photos. Interpretation of such aerial photos compared to current orthophotomaps

can enable the assessment of whether the floodplain vegetation community has been impacted.

Fronhas dam was built in 1985 and is located in the Alba River (Arganil, Coimbra, Portugal) (Fig. 1).

The River basin has an area of 652 km2 and has a reservoir area of 5350×1000 m

2 with a gross

capacity of 62100×1000 m3, an effective storage of 42500×1000 m

3 and a normal water level of 136

m. The dam body has a volume of 103×1000 m3 and has a maximum discharge of 500 m

3.s

-1. The

reservoir is used for hydropower generation and recreation. Touvedo dam was built in 1993 and is

located in Lima River (Viana do Castelo, Portugal) (Fig. 1). Its 134 m crowning in length has 43 m

height. The reservoir has a total volume of 15.5 hm3. The flood spillway, located in the dam body and

controlled by three segment gates was designed for a flow of 3200 m3.s

-1. The dam is used for

hydropower generation and produces about 67 GWh.year-1

. Vilarinho dam, inaugurated in May 21st

,

1972, is located in São João do Campo, Terras de Bouro (Braga, Portugal) (Fig. 1) and is part of the

River basin of Cávado River. Powered by Homem River, has a height of 94 m and a crowning of 385

m. The reservoir has a capacity of 118 hm3 and an area of 346 ha. It has a maximum discharge

capacity of 280 m3.s

-1. The reservoir is used for hydropower generation.

3.2. Methodology

In order to perform the georeferencing of the old black and white aerial photos obtained from the

Portuguese Geographic Institute (IGP) and current orthophotomaps (Reference System

PT-TM06/ETRS89) geographic information system tools were used. Georeferencing was performed

using Control Points (CPs) (e.g. houses, bridges, roads) sparsely distributed in the image. Once

obtained the georeferenced images, lateral limits of the riparian zone will be defined for both

riverbanks. Polygons of homogeneous strata of riparian vegetation – woody riparian vegetation – will

be delineated. Landscape metrics related with the spatial configuration, isolation, inter-connectivity,

and percentage of occupation of woody riparian vegetation will be calculated. Land use classes (e.g.

Agroforestry, Forestry, Agriculture and Urban) will be considered and evaluated in terms of

percentage of area occupied.

Projeto OASIS


Fig. 1. Location of the study areas.

IV. Results

4.1. Defining Reference Coordinate System

Reference Coordinate System was firstly defined based on the reference system used in the

orthophotomaps from the Portuguese Geographic Institute (IGP) (PT-TM06/ETRS89).

4.2. Georeferencing images

The concept of georeferencing involves establishing relationships between each point of the land and

its corresponding representation in the image through the assignment of coordinates linked to the

land and related to a spatial reference system. In order to georeference old aerial photos, it is

necessary to select control points (CPs) in the image, recognize them easily in the ground and/or in

other georeferenced images and establish their coordinates in a specific reference frame. Taking into

account the CPs, adjustment equations are calculated. These equations transform the image

coordinates (row, column) into coordinates (x, y) of the selected reference system. The procedure is

considerably simplified when it is possible to take coordinates from another image already

georeferenced. In consequence, CPs can be localized not only in vertices of polygons, intersections of

Lima River

Touvedo

(Salvador)

Touvedo

Vilarinho das Furnas

Fronhas

Campo do Gerês

Fronhas

N

Po

rtu

ga

l

Atl

an

tic

Oce

an

Projeto OASIS


roads, intersections of roads with railways or intersections of roads with rivers, with bridges, houses,

etc. In order to select the CPs, it has to be taken into account that the quality of the adjustment

obtained by the application of the functions of coordinates’ transformation depends on the quality of

the selected points, their correct localization and the distribution they have in the image. In this way,

the quantity of required points varies in function of the size of the image and the topographic

characteristics of the covering zone.

For the application of the proposed procedure common features e.g. houses, bridges, roads,

intersection points were identified on the old aerial photos and on the other datasets (reference

data). This was done with visual comparison of the old aerial photos with the reference data

(orthophotomaps or current satellite images) and using the tools available in ArcMap’s

Georeferencing Toolbar. Satellite images already georeferenced were used as a base to georeference

old aerial photos. CPs in old aerial photos were identified and assigned their real-world coordinates.

Under Georeferencing, using the tool “Add control points”, control points were added to the image.

From a mathematical point of view, an adjustment of a 1st

degree equation requires a minimum of 3

points, a second-degree equation requires a minimum of 6 points and, a third degree equation a

minimum of 10 points. Since CPs are just used for validation of the polynomial model that is

associated to the georeferencing process, the accuracy behavior exclusively depends on those points

and not on the polynomial order. In contrast with all the higher order polynomials (non-linear

distortion), the parameters of a 1st

order transformation possess a physical meaning (linear

distortion).

In this present work, a minimum of four control points were added, an adjustment of a 1st

order

transformation was made and a total RMS error ≤10 or ≈10 was considered (Fig. 2). However, more

CPs should be added in order to sharpen georeferencing of old aerial photos (ongoing process). On

the other hand, it is shown in Fig. 2 that the old aerial photo of Touvedo doesn’t match completely

with the current satellite image, after georeferenced. An attempt to solve this problem, is to split the

old aerial photo into two (split in the part where at least both images match) and georeference both

again (ongoing process).

Along the georeferencing process, in addition to the previously mentioned, some other difficulties

were found mainly due to the difficulty in detecting some features that could be used as CPs

associated to the more or less complexity of the land use (Fig. 3). This is a difficult, tedious and time

consuming process since the difference between the acquisition times of the old aerial photos and

the current satellite images is over 40 years. Although, there were few points that could be reliably

located on the old aerial photos, even though during this period of time the natural and human

Projeto OASIS


environment of the study areas changed more or less significantly. The old aerial photo of Touvedo is

being the easier one to work with, due to the minimal changes observed between the past and the

present (coverage by vegetation and human occupation). On the other hand, the addition of CPs in

the old aerial photos of Vilarinho das Furnas is hampered by the almost nonexistent of identifiable

features and density of the vegetation from the riverbanks to the land inside. This fact also lays a

problem in separating the riparian gallery from the remaining vegetation. Moreover, current satellite

image from Vilarinho das Furnas has a high amount of clouds that makes difficult the identification of

any feature on the ground that could be used as a CP and even the recognition of the river and

riparian vegetation. In the case of Fronhas, some of the enunciated problems persisted during the

georeferencing process.

Fig. 2. Example of the georeferencing process (Touvedo).

4.3. Images resampling

Image resampling is a process used to interpolate the new cell values of a raster image during a

resizing operation. There are many resampling methods available (bicubic, bilinear, cubic

Projeto OASIS


convolution, etc.), however Nearest Neighbor is one of the most commonly used resampling

methods. Each resampling method has strengths and weaknesses which should be considered

carefully. Nearest Neighbor (NN) resampling method functions by matching a pixel from the original

image to its corresponding position in the resized image. If no corresponding pixel is available, the

pixel nearest is used instead. NN works well with horizontal or vertical lines, but introduces

noticeable error along other linear features where pixel realignment is obvious, and for that reason is

generally considered the least accurate method (Brovelli and Minghini, 2012). NN remains widely

used because of the speed of implementation and simplicity. The NN method does not alter the

value of the input cells and should be used when resampling categorical data, such as surficial

geology, imagery or land cover. For this reason, this was the resampling method adopted to be used

in resampling the images (ongoing process).

Fig. 3. Comparison of old aerial photos (1960s) and current satellite images.

4.4. Structure of riparian vegetation and land use

Projeto OASIS


To analyse the structure of the riparian gallery – woody riparian vegetation – and assess the land use,

some steps must be followed in both old aerial photos and current satellite images.

i. Rivers reaches under study will be firstly divided into 300 m long sections (sampling units);

ii. Lateral limits of the riparian zone will be then manually digitalized for both riverbanks;

iii. Polygons of homogeneous strata of woody riparian vegetation – riparian patches – will be

delineated. This process will be done using visual screening of image features, namely the spatial

variation in pixel intensity pattern and the local contrast.

iv. Percentage of woody riparian vegetation will be calculated within each sampling unit.

v. Landscape metrics (e.g. NP – Number of Patches, MPS – Mean patch Size, PSCV – Patch Size

Coefficient of variation, MSI – Mean Shape Index, MPFD – Mean patch Fractal Dimension, MNN –

Mean Nearest Neighbor Distance, Mean Proximity Index, IJI – Interspersion and Juxtaposition

Index) within each sampling unit will calculated.

vi. In addition, land use will be assessed according to the following classes: e.g. Agroforestry,

Forestry, Agriculture and Urban. Patches of land use will be delimited for each buffer within each

sampling unit. Land use classes will be assessed in terms of percentage of area occupied, after

grouping the patches of the same class.

Note: These steps must be followed taking into account the same scale. All these processes are being

performed using the platform ArcGIS (ArcMap version 9.3) and metrics will be calculated using the

Patch Analyst extension.

These steps performed in both old aerial photos and recent satellite images will allow to make an

evaluation of the evolution of the riparian vegetation after the construction of the dam and land use

in both situations (i.e. prior and after dam construction). As previously referred much of the work is

still being developed and many of the methodologies adopted are being refined. Also noteworthy

and given the complexity of the work, the results presented should be considered preliminary.

V. Supportive literature

Aguiar, F.C., Fernandes, M.R., Ferreira, M.T. 2011. Riparian vegetation metrics as tools for guiding

ecological restoration in riverscapes. Knowl. Manag. Aquatic Ecosyst. 402, 21. DOI:

10.1051/kmae/2011074.

Aguiar, F.C., Ferreira, M.T. 2005. Human-disturbed landscapes: effects on composition and integrity

of riparian woody vegetation in the Tagus River basin, Portugal. Environ. Conser. 32(1), 30-41.

Projeto OASIS


Fernandes, M.R., Aguiar, F.C., ferreira, M.T. 2011. Assessing riparian vegetation structure and the

influence of land use using landscape metrics and geostatistical tools. Landsc. Urban Plan. 99,

166-177.

Fernandes, M.R., Aguiar, F.C., Ferreira, M.T., Pereira, J.M.C. 2013. Spectral separability of riparian

forests from small and medium-sized rivers across a latitudinal gradient using multispectral

imagery. Int. J. Remote Sens. 33(7), 2375-2401.

Ferreira, M.T., Aguiar, F.C., Nogueira, C. 2005. Changes in riparian woods over space and time:

Influence of environment and land use. For. Ecol. Manag. 212, 145-159.

Hillier, A. 2007. ArcGIS 9.3 manual. University of Pennsylvania, School of Design, Cartographic

Modeling Lab.

McGarigal, K., Cushman, S.A., Ene, E. 2012. FRAGSTATS v4: Spatial Pattern Analysis Program for

Categorical and Continuous Maps. Computer software program produced by the authors at

the University of Massachusetts, Amherst. Available at the following web site:

http://www.umass.edu/landeco/research/fragstats/fragstats.html

VI. References

Alldredge, B., Moore, G. 2012. Assessment of riparian vegetation sensitivity to river hydrology

downstream of a major Texas dam. River Res. Appl. DOI: 10.1002/rra.2625.

Brovelli, M.A., Minghini, M. 2012. Georeferencing old maps: a polynomial-based approach for Como

historical cadastres. e-Perimetron 7(3), 97-110.

Kattelmann, R., Embury, M. 1996. Riparian Areas and Wetlands. Sierra Nevada Ecosystem Project:

Final Report to Congress, Vol. III. Assessments and scientific basis for management options.

University of California, Centers for Water and Wildland Resources.

Naiman, R.J., Décamps, H., Pollock, M. 1993. The role of riparian corridors in maintaining regional

biodiversity. Ecol. Appl. 3(2), 209-212.

Salo, J., Kalliola, R., Häkkinen, Y., Niemelä, P., Puhakka, M., Coley, P.D. 1986. River dynamics and the

diversity of Amazon lowland forest. Nature 322, 254-258.

Swanson, F.J., Kratz, T.K., Caine, N., Woodmansee, R.G. 1988. Landform effects on ecosystem

patterns and processes. BioScience 38, 92-98.

Projeto OASIS


VI. Acknowledgements

The work is being supervised by the PI of the project OASIS – Como gerir rios regulados nas regiões

semiáridas? (PTDC/AAC-AMB/120197/2010), Francisca Constança Frutuoso de Aguiar (Invited

Scientist).

Special thanks to Francisca Aguiar and Maria do Rosário Fernandes for their support in the

development of the present work.

Documents

Universidade de Lisboa Faculdade de Ciências Departamento de … · Universidade de Lisboa Faculdade de Ciências Departamento de Biologia Animal Técnicas e Metodologias de Investigação