Analysis of Patch Geometry Characteristics in the Ridge Slough Patterned Landscape in the Everglades

Jing Yuan1, Joseph Delesantro1, Stephen Casey2, Sergio Padilla Paz3, Danielle Watts1, Matthew J. Cohen1

1Ecohydrology Laboratory; 2Environmental Sciences; 3Wildlife Ecology and Conservation, University of Florida, Gainesville, FL, USA

Results - Size and elevation

Methods Introduction

Fig.6 Vegetation maps of 4x4km sampling block. Ridge is grey slough is white. a.Conserved b.Dry c.Impounded

Fig.7 The three size classes of ridges showed no statistical difference in mean peat elevation.

Results - Size and anisotropy

Results - Length and width distribution

Our results suggest • Elevations independent of spatial organization, supporting the point-

scale conceptual model. • Tentative support for scale dependent anisotropy, but further work is

needed (size and anisotropy). • Preliminary support for ridge lengths following power-law scaling. Together, these results provide landscape geometric properties that can be used to evaluate simulation model performance of landscape pattern genesis.

Conclusion

• Continue these analyses to span hydrological conditions on more sampling block across WCA 3A.

• Test whether power law distribution of patch size can serve as a earlier warning indicator of landscape degradation.

Future directions

Watts, D., Cohen,M. Heffernan, J. , and Osborne, T. 2010. Ecosystems 13:813-827 Wu, Y., N. Wang, and K. Rutchey. 2006. Ecological Complexity 3:183-192. Cohen, M., Watts, D. Heffernan, J. and Osborne, T. 2011.Critical Reviews in Environmental Science and Technology 41:395-429 Nungesser, M. 2011 Wetlands Ecology and Management 19:475-493. Larsen, L.,Harvey, J. and Crimaldi, J. 2007,Ecological Monographs 77(4):591-614. Scanlon, T., K. Caylor, S. Levin, and I. Rodriguez-Iturbe. 2007. Nature 449:209-U204.

Reference

Ridge Elevation

Ridge Size Class

Pea

t Ele

vatio

n (c

m)

190

200

210

220

230

240

250

Small Medium Large

y = 65.321x-1.555 R² = 0.9587

0.001

0.01

0.1

1

10

20 200

y = 64.444x-1.398 R² = 0.9684

0.001

0.01

0.1

1

10

20 200

y = 62.689x-1.539 R² = 0.9748

0.001

0.01

0.1

1

10

20 200

y = 33.65x-1.189 R² = 0.9886

0.001

0.01

0.1

1

10

20 200 2000

y = 32.634x-1.196 R² = 0.9591

0.001

0.01

0.1

1

10

20 200 2000

y = 18.721x-0.914 R² = 0.968

0.001

0.01

0.1

1

10

20 200 2000

a.

b.

c.

Data process • Binary maps were created designating sloughs as 0’s and ridges as 1’s. • Ridges were then grouped according to the whether ridge cells are adjacent to one

another and various statistics were gathered for each plot • The number of cells belonging to each ridge was summed across individual rows and

columns to generate length and width statistics

Fig.8 (Upper panel)Log transformed ridge patch average length to average width ratio as a function of patch size across hydrological conditions.(Lower panel )Ridge patch orientation as a function of patch size. Orientation is the angle formed by the major axis of the ellipse which has the same 2nd moment of inertia as the ridge patch (0 = in direction of flow, -/+90=perpendicular to flow).Both metrics support hypothesis 2 indicating that both patch elongation and alignment in the direction of flow become greater with increasing patch area. However, the impounded site shows a weaker response in patch elongation.

Size classes Small 1-1000m2

Medium 1001-10000m2

Large 10001-100000m2

This supports our hypothesis that elevations are independent of size (1.)

-100

-80

-60

-40

-20

0

20

40

60

80

100

400 4000 40000 400000

Orie

ntat

ion

(deg

rees

fr

om v

ertic

le)

-100

-80

-60

-40

-20

0

20

40

60

80

100

400 4000 40000 400000 4000000

Conserved Impounded Dry

y = 0.0238ln(x) - 0.0244 R² = 0.0403

-0.5

0

0.5

1

400 4000 40000 400000

y = 0.0568ln(x) - 0.2312 R² = 0.2215

-0.5

0

0.5

1

400 4000 40000 400000 4000000

y = 0.0589ln(x) - 0.2685 R² = 0.1894

-0.5

0

0.5

1

400 40000 4000000

Red square is 4 ×4 km sampling block. From north to south is hydrologically dry, conserved and impounded

Impounded Conserved Dry

Ridge width(m) Ridge width(m) Ridge width(m)

Ridge length(m) Ridge length(m) Ridge length(m)

Patch size (m2) Patch size (m2) Patch size (m2)

Log(

leng

th/w

idth

)

P(W

≥ w

) P(

L ≥

l)

• Everglades ridge slough patterning has several striking attributes: 1. Patches are oriented parallel to historical flow, with elongated

ridges embedded within a matrix of sloughs, and 2. Ridges occupy higher elevation sites than sloughs (ca. 25 cm).

• These attributes fail to capture the apparent fractal nature of patch geometry, where ridges vary widely in size, and possess substantially convoluted margins.

• Our objective was to investigate pattern geometry, with a focus ridge patches as geometrically regulating landscape discharge.

Fig.1 Ridge and slough landscape , March 2012. Photo credits: Robert Sobczak (www.gohydrology.org)

Hypotheses

1.The point scale model dictates peat elevation is purely a function of local hydrology ,unaffected by ridge size. Ridge elevations should be independent of mechanisms controlling patch size

2. Larger ridges have a greater probability of occluding flow and as a result, should be preferentially acted upon by the discharge competence mechanism, causing ridge elongation and orientation in the direction of flow to increase with ridge size.

1. 2. 3.

3.While the ridge slough landscape has scale free attributes such as power law scaling in areas , it also appears to show scale dependent characteristics such as regular patterning. To reconcile this we propose that the distal negative feedback only acts in the width dimension, allowing power law distributions in the length dimension and regular patterning laterally.

Ridge

Ridge Lateral expansion

Hydroperiod & water depth

Soil accretion

Landscape discharge competence

+

+

-

-

-

-

+

-

Fig.4Conceptual model of the discharge competence mechanism. (Cohen et al.2011)

Fig.3 Illustration of ridge patch expansion

Fig.2 Ridge and slough conceptual model at point scale (Watts et al. 2010).

Fig.5 Research site map

1. Initial random propagation of ridge patches 2. Lateral expansion leads to inundation and stress to ridges 3. Ridges expand longitudinally with little impact on flow

Fig.9 Cumulative distributions of average widths (w) and lengths (l). Both length and width appear to follow power law distributions. Preliminary results show larger ridges in the conserved and dry sites as being composed of groups of many smaller interconnected ridges. This inter-ridge connectedness may allow for the coexistence of scale-free distributions and scale-dependent distal negative feedbacks. The slope of the distribution appears to correspond to hydrologic conditions, with more inundated sites having steeper slopes and smaller ranges.

-100

-80

-60

-40

-20

0

20

40

60

80

100

400 4000 40000 400000 4000000