Assessing ecosystem vulnerability

Assessing climate change vulnerability of ecosystems in the Southeast

Jennifer CostanzaJanuary 15, 2015

Vulnerability assessments: a key step in adaptation planning

USFWS 2010 Climate Change Strategy Objective: Develop Expertise In and Conduct Adaptation Planning for Key Species and Habitats

National Fish, Wildlife, and Plants Climate Adaptation Strategy 2012:Vulnerability assessments are important for management planning and decision-making under uncertainty.

Glick et al. 2011 Scanning the Conservation Horizon:Vulnerability assessments are the first steps to facilitating ecosystem resistance and resilience to climate change.

Climate change vulnerability as part of management and conservation

Glick et al. 2011

Assessing Climate Change Vulnerability

Glick et al. 2011

Objective

Assess climate change vulnerability for a set of ecosystems in the Southeast and Caribbean.

• Phase 1: Synthesize literature and GIS data on sensitivity, exposure, and adaptive capacity for 12 ecosystems

• Phase 2: Use NatureServe’s Habitat Climate Change Vulnerability Index (HCCVI) for 2 of these ecosystems

Study area

Approach: Ecosystem selection

Initial criteria: likely to be sensitive to climate change• Coastal or coastal plain regions• Wetlands• High elevation systems• Endemic to geographically discrete conditions, special soils

or other unique environments• IUCN Threatened• Suggestions from LCCs

Resulted in a list of 55 potential ecosystems in the Southeast and Caribbean.

Approach: Ecosystem selection

Ecosystem (NatureServe Ecological System)

Caribbean Coastal Mangrove

Caribbean Montane Wet Elfin Forest

Central Atlantic Coastal Plain Wet Longleaf Pine Savanna and Flatwoods

Central Florida Wet Prairie and Herbaceous Seep

East Gulf Coastal Plain Near-coast Pine Flatwoods

East Gulf Coastal Plain Southern Loess Bluff Forest

Edwards Plateau Limestone Shrubland

Edwards Plateau Mesic Canyon

Nashville Basin Limestone Glade and Woodland

South-Central Interior Mesophytic Forest

Southern Coastal Plain Nonriverine Cypress Dome

Southern Coastal Plain Seepage Swamp and Baygall

Many vulnerability assessments have been done for species

Vulnerability assessments for ecosystems or habitats are more challenging.

St. Mark’s National Wildlife Refuge. Credit: Alan Cressler

Cedar glade, Rutherford Co., TN, Credit: Alan Cressler

Vulnerability assessments for ecosystems or habitats are more challenging.

Phase 1 assessment approach

Climate factors that affect:• Distribution of the ecosystem• Important ecological processes • Dominant plant species• Plant and animal species of conservation importance

Focus on:• Non-climate stressors• Ecosystem function• Connectivity• Environmental/topographic complexity• Protection status and management

Qualitative, relative vulnerability rating

Phase 1 major GIS data

Sensitivity• WorldClim 1950-1999 data: seasonal max and min temperature, total precipitation

Exposure• Downscaled climate projections (K. Hayhoe et al.), averaged across GCMs for

A1FI and B1 scenarios, 2040-2060, 2080-2100• NOAA sea level rise: recent data and projections for 1 ft and 6 ft rise • Sea level rise vegetation impacts from SLAMM modeling

Adaptive Capacity• Current ecosystem total area (GAP land cover)• Mean patch size (GAP land cover)• Proportion under conservation protection (US Protected Areas Database)• Proximity to urban areas (GAP land cover)• Projected urbanization (Terando et al. 2014)• Human modification (Theobald et al. 2012)• Elevation heterogeneity (National Elevation Database)

Phase 1: literature and other sources

• Scientific literature• Reports• Wildlife Action Plans• US Forest Service Tree Atlas for dominant species

Especially for: • Climate sensitivities• Non-climate threats like invasive species• Management concerns

Phase 1 Example results forEast Gulf Coastal Plain Near-coast Pine Flatwoods

Phase 1 SensitivityEast Gulf Coastal Plain Near-coast Pine Flatwoods

14

16

18

20

22

24

26

28

30

32

34

36

Winter Spring Summer Fall

Max

Tem

p (

˚ C

)

0

5

10

15

20

25

Winter Spring Summer Fall

Min

Tem

p (

˚ C

)

Dominant species (longleaf, slash pine, wiregrass) have low sensitivities here.Fires and hurricanes are important and are sensitive to climate.Some rare salamanders and other species depend on mesic, periodically saturated soils.The ecosystem is also sensitive to sea level rise.

Phase 1 ExposureEast Gulf Coastal Plain Near-coast Pine Flatwoods

● ●

●

●●

●

●

● ●

●

●

● ●

●

●

●

0

2

4

6

Winter Spring Summer FallSeason

Cha

ng

e in

tem

p.

(°C

)

Scenario, time period●

●

●

●

A1FI, mid−centuryB1, mid−centuryA1FI, end of centuryB1, end of century

● ●

●

●

●

●

●

●

●

●

●

●●

●

●

●

0

2

4

6

8


Cha

ng

e in

tem

p.

(°C

)


●

●

●


●●

●

● ● ●●

● ●●● ●

●

●

●

●

−50

0

50

100


Ch

an

ge in p

recip

. (%

)


●

●

●


Max temp. Min temp.

Precipitation

●● ● ●

●

●

●

● ●

●

●

●

●●

●

●

0.5

1.0

1.5

2.0


Ch

an

ge in s

td.

dev.


●

●

●


● ●● ●

●

●

●

●

●

● ●●

●●

●

●

0.8

1.2

1.6


Ch

an

ge in s

td.

dev.


●

●

●


● ●● ●

●

●● ●

● ●●●

●

●

●●

0.5

1.0

1.5

2.0


Ch

an

ge in s

td.

dev.


●

●

●


Max temp. Min temp.

Precipitation



For most seasons, time periods and scenarios, conditions could become hotter and drier.

The water table could become lower, especially during dry periods (Lu et al. 2009).

Wildfires could become more common or more intense, but there is uncertainty and debate about this.

Between 1% and 8% of the ecosystem may be directly inundated by sea level rise, and 2-3% could change vegetation type.

Hurricane intensity is projected to increase, leading to overstory tree mortality, especially on older, unburned sites.

Longleaf and slash pine are not projected to change in importance (US Forest Service Tree Atlas).

Phase 1 Adaptive CapacityEast Gulf Coastal Plain Near-coast Pine Flatwoods

Recent fire suppression is a major threat.

Relatively small average patch size (0.9 ha)

Occurs fairly close to development, likely placing more constraints on management.

6% projected to be developed by 2050, 12% by 2100

Invasive species such as cogongrass are problems.

But…Species in this system occur in other adjacent systems.

High degree of protection: over 50% is protected currently.

0%

10%

20%

30%

40%

50%

60%

Pro

po

rtio

n o

f ec

olo

gica

l sys

tem

ext

ent

Protected Developed Developed

2050 2100

Phase 1 Overall vulnerabilityEast Gulf Coastal Plain Near-coast Pine Flatwoods

Moderate

Some sensitivity and exposure, especially to changes in fires, hurricanes, hydrology, and sea level.

Substantial adaptive capacity.

Phase 1 Example results forNashville Basin Limestone Glade and Woodland

Flat Rock Cedar Glades and Barrens, TN, Credit: Alan Cressler

Phase 1 Sensitivity, Exposure, Adaptive CapacityNashville Basin Limestone Glade and Woodland

Sensitivity:• Seasonal extremes in moisture, and longer term drought, often in late summer and fall• Many plant species tolerate abiotic stress• Woody invasion into grasslands slows under drought

Exposure:• Temperature increases, especially minimum temps.• Precipitation projections more uncertain• Could lead to more droughts• Changes in hydrology uncertain

Adaptive Capacity:• Limited geographic distribution and restricted to unique edaphic conditions• Threatened by conversion to human land uses, especially in

Nashville-Murfreesboro area

Vulnerability: High due to low adaptive capacity and potential changes in drought and hydrology.

Phase 2 Assessment Approach: HCCVI for two ecosystems

Adapted from Comer et al. 2012

Phase 2: HCCVI

Two ecosystems:East Gulf Coastal Plain Near-Coast Pine FlatwoodsNashville Basin Limestone Glade and Woodland

Two time periods: Mid-century (c. 2050)Late century (c. 2100)

Minimum and maximum values for each time period when applicablefrom alternative emissions scenarios or other uncertainty.

Phase 2: Dynamic process alterationand dynamic process forecast

LANDFIRE State-and-transition model for longleaf pine ecosystem


Current: Dynamic process alteration

Future: Dynamic process forecast


Dynamic process

alteration

Dynamic process forecast

Pine Mid C. min. 0.68 0.31

Flatwoods Mid C. max. 0.68 0.31

Late C. min. 0.68 0.31

Late C. max. 0.68 0.29

Nashville Mid C. min. 0.49 0.22

Basin Mid C. max. 0.49 0.22

Late C. min. 0.49 0.21

Late C. max. 0.49 0.21

Phase 2: Calculation of climate stress index

The degree to which climate will change in the future in an ecosystem’s range.

Compared recent (1980-1999) climate data (Maurer et al. 2002)with projected data (K. Hayhoe) for each ecosystem.

Monthly minimum and maximum temperature, precipitation(36 variables)

For each variable at each climate data pixel:Does the projected mean value exceed the recent value +/- 2 stdev.?

Climate stress index: 1 – proportion of variables that exceed recent values

Average across pixels in the ecosystem.

East Gulf Coastal PlainNear-Coast Pine Flatwoods

Nashville Basin Limestone Glade and Woodland

Climate stress index

Pine Mid C. min. 0.85

Flatwoods Mid C. max. 0.75

Late C. min. 0.75

Late C. max. 0.55

Nashville Mid C. min. 0.88

Basin Mid C. max. 0.78

Late C. min. 0.77

Late C. max. 0.51

Phase 2: Calculation of envelope shift index

The degree to which the suitable climate envelope will change.

Define the current climate envelope using MaxEnt and map to the current landscape.

Project where that envelope will occur in the future under alternate scenarios.

Envelope shift index: Average suitability score for the ecosystem’s current extent.

East Gulf Coastal Plain Near-Coast Pine Flatwoods

ObservedB1

A2A1FI

Mid-century

East Gulf Coastal Plain Near-Coast Pine Flatwoods

ObservedB1

A2A1FI

End of century

Envelope shift index

Pine Mid C. min. 0.81

Flatwoods Mid C. max. 0.63

Late C. min. 0.54

Late C. max. 0.25

Nashville Mid C. min. NA

Basin Mid C. max. NA

Late C. min. NA

Late C. max. NA

System Period Sensitivity Indirect Effects

Adaptive Capacity

Resilience Vulnerability

Pine Mid C. min. 0.74 0.79 0.50 0.64 Low

Flatwoods Mid C. max. 0.65 0.77 0.42 0.60 Moderate

Late C. min. 0.65 0.79 0.42 0.61 Moderate

Late C. max. 0.50 0.77 0.37 0.57 Moderate

Nashville Mid C. min. 0.55 0.70 0.45 0.57 Moderate

Basin Mid C. max. 0.49 0.66 0.37 0.52 High

Late C. min. 0.49 0.70 0.37 0.53 High

Late C. max. 0.36 0.66 0.32 0.49 Very High

Phase 2: HCCVI Vulnerability Scores

●

●

●

●

Moderate

Moderate

Low Low

Moderate

High

Moderate

High

Very high

0.00

0.25

0.50

0.75

1.00

0.000.250.500.751.00

Resilience

Se

nsitiv

ity

System and time period●

●

●

●

Pine Flatwoods mid−century min

Pine Flatwoods mid−century max

Pine Flatwoods late−century min

Pine Flatwoods late−century max

Nashville Basin mid−century min

Nashville Basin mid−century max

Nashville Basin late−century min

Nashville Basin late−century max

Phase 2: HCCVI Vulnerability Scores

Assessing vulnerability:Summary and future needs

Both approaches gave relatively similar vulnerability ratings for the two ecosystems

Each approach has advantages:• Phase 1 approach was able to incorporate a wider variety of information and provide

more nuanced assessment• Phase 2 approach produced quantitative scores that could be compared to one another

Need to incorporate:• Species interactions• Feedbacks and interactions among vulnerability components

Need to consider:• Individual responses to climate change: ecosystems will not likely change as a unit• Assessing vulnerability for ecosystems is more complex than for species• What is the desired future condition? • Focus on major species and ecological processes?

Sources of uncertainty and future research needs

• Potential change in precipitation• Effects of climate on cloud cover (for Caribbean Elfin

Woodlands)• Impacts on hydrologic regime• Impacts on disturbance regimes• Future of biotic interactions, tropic relationships, dispersal

under climate change

Assessing vulnerability will be important to do for a wider number of ecosystems in the Southeast.

Thank you!

Collaborators:Jaime Collazo, Matthew Rubino, NC State, Coop. Fish and Wildlife Research UnitScott Beck, NC StateMilo Pyne, Rickie White, NatureServeAdam Terando, USGS Southeast Climate Science Center

Funding:USGS Southeast Climate Science CenterUSGS Gap Analysis Program

Contact me with questions or for a copy of our draft report: [email protected]

Assessing ecosystem vulnerability

Environment

climate changecoastal

key species

set of ecosystems

species responses

potential ecosystems

adaptation planningusfws

conduct adaptation planning

ecosystem resistance