Novel Computational Tools to Analyse Fragmented Forests Vronique - - PowerPoint PPT Presentation

Novel Computational Tools to Analyse Fragmented Forests

Véronique Lefebvre, Marion Pfeifer, Andrew Bradley and Robert Ewers

ATBC 2013

Extracting fragments and their properties from maps

• Fragmentation of forest affects biodiversity
• To find out how we can:

– Define what fragments are and delineate them on maps obtained from satellite images – Estimate potential biodiversity drivers for each fragment (size, shape, connectivity..)

• We present novel image processing based

methods to

– Delineate fragments on maps – Estimate fragments properties

• Novel methods can make use of prior

knowledge on studied species and local area

• And are designed to cope with resolution and

pixel geometry constraints of raster maps Distinct fragments? Similar “shape”?

Part 1: Patch delineation by CCL

‘on’ pixel ‘off’ pixel Habitat binary map Connected component labelling Patch 1 Patch 2

Common technique: Connected Component Labelling (CCL) Problems:

Does not represent species perception of landscape

• Nor experimenters’

perception of landscape cannot use prior knowledge very sensitive to forest classification

10 km

Part 1: Patch delineation by new method

• Our method uses

ecological knowledge

• To disconnect

weakly connected chunks of pixels Delineation reflects species perception of landscape definition of patches is adaptable

Connected component labelling Patch 1 Patch 2 New delineation method 3 patches

Effective fragment – Concept and definition

• Chunks of forest may be connected but not perceived as such

– A stretch of forest may be too narrow to be a corridor – Habitat suitability may vary with the distance to forest edge

How to decide where to “separate” connected pixels?

Concept of effective fragment

• To delineate effective fragments the method uses:

– the Minimum corridor width (MCW) to find weak links – the Depth of Edge Influence (DEI) to find core area

• MCW and DEI can be obtained from species abundance data, local knowledge and

literature, and map classification confidence

Delineation technique

1) Find cores and corridors from the distance map 2) Find where to cut all weak links (narrower than MCW) with the watershed segmentation 3) Reconnect edge chunks (less then DEI) to most strongly connected core

• Can incorporate matrix

element, e.g. water, pastures, urban, by adding weights to the distance map

Landscape map Distance map MCW / DEI Watershed segmentation Reconnection

Part 1: Patch delineation – Example Result

Comoros Islands Forest

Binary map of landscape and measurement locations

Connected component labelling New delineation method

1 km Watershed delineation landscape segmented into ecologically meaningful fragments

Part 2 – Fragment characteristics

• To compare fragments between each other we can extract their geometrical

characteristics from simple binary maps and fragments delineation:

– Area – Core area – Potential dispersal area from a fragment – And shape Straightforward from patch delineation Forest map Potential dispersal area

?

Shape descriptors

Compactness Smoothness

Compactness and contour smoothness can describe different types of habitat

Commonly used shape descriptors do not distinguish between these 2 properties of shape

Bogaert et al. 1999, Environmental and Ecological Statistics

Compact shape: effective in conserving resources Convoluted shapes: effective in enhancing interaction with the surroundings Smooth shapes: Higher resistance to disturbance

Compactness

How packed is the shape? Longest distance within fragment Circle of same area

The compactness measure shows the “spread” of the fragment compared to the most packed shape. Advantages: - measures only compactness

• does not use a perimeter estimation

Contour smoothness

How wriggly is the contour line of a shape? Suggestion: counting the number of indents in shape contour Method: Counting the number of zero crossing

• f the contour curve derivatives

Compute the proportion of smooth perimeter:

Advantage:

• Only describes smoothness

But it requires a perimeter estimation (which varies with resolution)

• Inspired by:

Bogaert et al. 1999, Environmental and Ecological Statistics

indents

Perimeter calculated using distances between mid-points of border pixels

Comparison of common and suggested shape descriptors

On the 10 biggest fragments

• f the Comoros forest

Effective fragments obtained by the Watershed method

1 2 3 4 5 6 7 8 9 10 distance from landscape top (km) distance from landscape left (km) 1 2 3 4 5 6 7 8 9 10 11 12 13 14 2 4 6 8 10 12 14

Comparison of common and suggested shape descriptors

# 2 # 3 # 1 # 4 # 5 # 6 # 9 # 8 # 10 # 7 # 8 # 10 # 6 # 9 # 4 # 3 # 2 # 7 # 5 # 1 # 3 # 4 # 2 # 1 # 9 # 6 # 5 # 10 # 8 # 7 # 5 # 1 # 7 # 2 # 4 # 10 # 3 # 9 # 6 # 8

Fragments ordered by:

# 1 # 2 # 3 # 4 # 5 # 6 # 7 # 8 # 9 # 10

+

• Shape factor

(compactness) Compactness Fractal dimension Smoothness Area

• +

Shape factor and fractal dimension classifications mainly reflect area order

Variation of shape descriptors with area

Forest patches from several landscapes

10

1

10

2

10

3

0.2 0.4 0.6 0.8 1 10

1

10

2

10

3

5 10 15 20 25 30 10

1

10

2

10

3

0.2 0.4 0.6 0.8 1 10

1

10

2

10

3

1 1.5 2 2.5 3

Random patches Fragment area Fragment area

Compactness Compactness Smoothness Smoothness Fractal Dimension Fractal Dimension Shape Factor Shape Factor

Our metrics are less determined by area easier comparison between fragments

New descriptors New descriptors

Examples of randomly generated patches

Variation of shape descriptors with each other

Forest patches Random patches Compactness w.r.t smoothness (new) Shape factor w.r.t. fractal dimension

5 10 15 20 25 30 0.3 0.4 0.5 0.6 0.7 0.8 0.9

Our descriptors are not functions of each other reflect distinct shape properties Area

What can fragment delineation and descriptors do for forest fragmentation studies

compactness Biodiversity index

• Delineation and fragment descriptors can help selecting plot locations within several patches
• f widely different properties
• Fragment delineation is also useful in finding out fragments history Robert Ewers’ talk at

8:45 in this session Forest map and plot locations

Trend?

• Used to study biodiversity responses to fragments properties (area, potential dispersal area,

compactness, smoothness)

• But often not enough fragments are measured in a landscape

Thanks !

• To all researchers helping us to collect Biodiversity data for the BioFrag project:
• Thanks to the team !

– Andrew Bradley (the remote sensing pro) – Marion Pfeifer (the ecology reference who patiently teaches me everything) – Robert Ewers (the wise boss)

• Thanks for your attention
• The delineation method and metric code is available in Matlab with the image

processing toolbox. It can be recoded in another language or included in an existing software. We are open to collaborations

v.lefebvre@imperial.ac.uk

http://biofrag.wordpress.com/