1 1. Land cover: carbon is a key attribute L d b i k tt ib t - - PowerPoint PPT Presentation

1 L d b i k tt ib t 1. Land cover: carbon is a key attribute 2. Land use: carbon storage is altered constantly by policies and i management actions 3. Climate change: national/international policies require C assessments and mitigation actions 4. Locally/regionally, carbon assessment provides key indicators y g y y

• f ecosystem sustainability

Policy example y p

“Access to the U.S. Geological Survey’s ‘visualization tool’ to assess the amount of carbon absorbed by landscapes …”

Policy example y p

How is carbon related to land cover and land use? How is carbon related to land cover and land use?

CO2 A h i CO2 Atmosphere concentration Rates of contributing components

C b l i

Forests Wetlands

Carbon cycle in different land t cover types

Agricultural lands Grasslands Aquatic systems

Mapping land cover for C assessment: challenges challenges

Two examples: U.S. and Kenya Questions asked:

• What are the current land cover and land use?
• What are the current carbon stocks and

sequestration rates based on the current land cover and land use?

• What may be the future potential land cover and land

use?

• What may be the future potential carbon stocks and

sequestration rates under the future scenarios of sequestration rates under the future scenarios of land cover/land use?

Starting from a land cover map

1992 U.S. land cover

Develop past land cover change metrics

Select future land cover change scenarios Select future land cover change scenarios

IPCC climate change scenarios are tied directly to factors that control land cover/use change, such as: GDP, population, energy needs technology development environmental policies etc

IPCC AR4 SRES scenarios IPCC AR5 RCP scenarios

needs, technology development, environmental policies, etc.

Select future land cover change scenarios Select future land cover change scenarios

U.S. example Kenya example

IPCC AR4 SRES scenarios IPCC AR5 RCP scenarios

IPCC scenarios are global scale, downscaled to continental scale

We further downscaled the scenarios to a regional scale using land cover data

• Land cover composition is

derived for cells/pixels

• Then paired to the underline

factors supporting the i scenarios.

• For example, amount of

agricultural lands needed to support population growth

• Using this method, regional

land cover transition is developed developed

Used a state and transition model for LULC change modeling

• States

States – major land use/cover (LULC) classes.

– World Resources Institute LULC classification.

• Transitions

Transitions – changes between state classes. Att ib t Att ib t d fi d f t t d t iti t

• Attributes

Attributes – defined for states and transitions to track changes in ecosystem carbon by ecoregion S ti l d ti l bilit Spatial and non-spatial capability.

• Monte Carlo analysis to assess uncertainty.

y y

• Spatially explicit and annual time steps.

Results of downscaled scenarios

Using a landscape model to link the past land cover change metrics to the downscaled scenarios to produce change metrics to the downscaled scenarios to produce annual land cover maps from the past to the future

A1B 1992 land cover B1 A2 B1

As the result, we have annual maps between 1992 2050 1992-2050

1992 2050

A1B Scenario – Little Rock, Arkansas

Little Rock Pine Bluff

Other projections can be similarly produced

2006 Age 2050 Age F t A Y Forest Age - Years

250

Using biogeochemical g g models and the land cover projections, we p j , then estimated both the present and future p carbon stocks and sequestration rates q

Linking C assessment results to other ecosystem services

• Example for a test in

a county in Mississippi

• Using “Ecosystem

g y Service Change Indicator” method

Kenya Example

(Proof of Concept)

Kenya proof of concept for climate change mitigation by increasing C sequestration

• Downscale global environmental change

scenarios to local and regional scales useful for decision/policy makers.

Sensitivity analysis for mitigation/adaptation – Sensitivity analysis for mitigation/adaptation

• Focus on changes in land use as primary driver

g p y

• f ecosystem carbon dynamics.

Select future land cover change scenarios Select future land cover change scenarios

U.S. example Kenya example

IPCC AR4 SRES scenarios IPCC AR5 RCP scenarios

RCP85 MESSAGE Future Scenario Model, 2005‐2100 Transition 'cu' (Cropland to Urban)

0.000% ‐ 0.008% 0 009% ‐ 0 028%

Transition cu (Cropland to Urban) Average Annual Percentage

• f Half‐Degree Grid Cell:

0.009% 0.028% 0.029% ‐ 0.058% 0.059% ‐ 0.121% 0.122% ‐ 0.231%

1,000 2,000 500 Kilometers

Predicted `sc` (Secondary to Cropland) by Scenario Average Annual Percentage of Half‐Degree Grid Cell

RCP Scenario Variability

0.000% 0.001% ‐ 0.020% 0.021% ‐ 0.100% 0.101% ‐ 1.000%

RCP 8.5

0.000% 0.001% ‐ 0.020% 0.021% ‐ 0.100% 0.101% ‐ 1.000%

RCP 6.0

1.001% ‐ 4.234% 1.001% ‐ 3.586% 0.000% 0.001% ‐ 0.020%

RCP 4.5

0.000% 0.001% ‐ 0.020%

RCP 2.6

0.021% ‐ 0.100% 0.101% ‐ 1.000% 1.001% ‐ 3.232% 0.021% ‐ 0.100% 0.101% ‐ 1.000% 1.001% ‐ 4.178%

Temporal Distribution of Land Cover Classes & transitions between 2000-2100 under RCP8.5 (km2)

Note: RCP8 5 used as reference conditions L d l T iti Note: RCP8.5 used as reference conditions Land cover classes Transitions

Temporal distribution of agricultural and forest carbon stock between 2000-2100 under the RCP8.5 reference scenario (km2)

0-9 yrs. >44 yrs. 10-44 yrs.

Under a Reference RCP8.5 Scenario

Forest Biomass Carbon Change Forest Area Forest Area 2000: 43,177 km2 2100: 34,447 km2 Carbon Stock Carbon Stock 2000: 365 TgC 2100: 240 TgC Carbon Density Carbon Density 2000: 8.4 Gg/km2 2100: 6.9 Gg/km2 Carbon Flux: +125 TgC Carbon Flux: +125 TgC

Mitigation scenarios applied to forest biomass carbon

• Strategy 1. Protection of ‘old growth’ forest.

– No harvest allowed on stands older than 40 yrs. Minimum age of 10 yrs for harvest – Minimum age of 10 yrs. for harvest.

• Strategy 2. Harvest reduction over baseline

– Reduction of harvest by 5%. R d ti f h t b 20% – Reduction of harvest by 20%.

• Strategy 3. Increase restoration efforts

– 5% increase. 20% increase – 20% increase

2000 2100

y-axis units = km2 Range = min/max over 25 simulations Line = mean RCP8.5 5% Combined Scenario 20% Combined Scenario

Forest Biomass Carbon Change by Mitigation Scenario

Forest Area Carbon Carbon Density (Mg Carbon Flux (km2) Stock (Tg C) y ( g C/km2) (Tg C) Scenario 2000 2100 2000 2100 2000 2100 Avg. Annual Total by 2100 Baseline Reference RCP85 43,177 34,447 365.4 240.3 8.5 7.0 1.251 125.1 All 5% Actions 43,143 43,428 365.6 264.5 8.5 6.1 1.011 101.1 All 20% Actions 43,296 65,939 366.0 374.4 8.5 5.7

• 0.084
• 8.4

Ecosystem Carbon Change by Mitigation Scenario

Carbon Stock by Scenario (2100)

Baseline Reference RCP85 300 350 400

Carbon Stock by Scenario (2100)

Baseline Reference RCP85 Old Growth Protection and 5% Reduction in Harvest Rate Old Growth Protection and 20% Reduction in Harvest Rate 5% Reduction in Deforestation 150 200 250 Tg C 5% Reduction in Deforestation Rate 20% Reduction in Deforestation Rate 5% Increase in Reforestation Rate 20% Increase in Reforestation 50 100 Rate All 5% Actions All 20% Actions

Carbon Stock Change Compared to Reference

B li R f RCP85 40 0% 50.0% 60.0%

Carbon Stock Change Compared to Reference

Baseline Reference RCP85 Old Growth Protection and 5% Reduction in Harvest Rate Old Growth Protection and 20% Reduction in Harvest Rate 5% Reduction in Deforestation 10 0% 20.0% 30.0% 40.0% Tg C 5% Reduction in Deforestation Rate 20% Reduction in Deforestation Rate 5% Increase in Reforestation Rate 20% Increase in Reforestation

• 10.0%

0.0% 10.0% 20% Increase in Reforestation Rate All 5% Actions All 20% Actions

Analysis: forest biomass carbon example

• Based on this proof of concept application:

Analysis: forest biomass carbon example

– RCP85 results in large scale demand for anthropogenic land use in Kenya and will likely result in a decline in Kenya forest biomass C t k stocks. – Mitigation actions aimed at forest harvest alone have minimal impact on national-scale biomass carbon dynamics impact on national scale biomass carbon dynamics. – Avoiding deforestation (20% decreases) and increasing reforestation (20% increases) have strong potential to increase ( ) g p biomass carbon in forests, especially at high levels of implementation. A i t t d i l t ti f iti ti t t i t th 20% – An integrated implementation of mitigation strategies at the 20% level has the most impact on increasing forest carbon biomass of all evaluated scenarios.

Important Caveats

• We did not have land cover history, we used downscaled

RCP 8.5 as a default.

• Stratification was based on WWF ecofloristic zones

W d IPCC LUCCF B t P ti Ti 1 d f lt

• We used IPCC LUCCF Best Practices Tier 1 default

values:

IPCC th t d t t i f t b b l – IPCC growth rates used to categorize forest carbon by age class (early (0-9 yrs.), mid (10-44 yrs.), late (>44 yrs.)). – Agriculture stocks extracted from CDIAC biomass spatially Agriculture stocks extracted from CDIAC biomass spatially explicit maps (average value by ecological region). – Fixed values for grass/shrub (IPCC shrubland category). g ( g y) – Mean soil carbon values extracted by LULC type and ecological region from ISRIC World Soil Database at 1-km resolution.

Uncertainties Uncertainties

• There are many!

M hi i d h d f d li l li LULC

• More sophisticated method of downscaling to localize LULC

changes.

• Climate assumed to be constant for LULC change (but dynamic in
• Climate assumed to be constant for LULC change (but dynamic in

carbon analysis)

• Many default values used for the Kenya example

y y p

• High degree of uncertainty in forest age initialization
• Natural disturbance regime not modeled

Natural disturbance regime not modeled.

• Initial/starting LULC conditions from remote sensing data range in

quality. q y

• Very limited Monte Carlo analysis implemented (100 simulations for

reference case; 25 simulations for mitigation scenarios)

In Summary …, we answered the following i (b i h )

What are the current land cover and land use?

questions (but with many many caveats)

• What are the current land cover and land use?
• What are the current carbon stocks and sequestration

rates based on the current land cover and land use?

• What may be the future potential land cover and land

y p use?

• What may be the future potential carbon stocks and
• What may be the future potential carbon stocks and

sequestration rates under the future scenarios of land cover/land use?

Parting Thoughts Parting Thoughts

• Applications directly link land cover mapping to

policies and land management

• They give purpose to why we are doing land

cover mapping cover mapping

• Maybe in the next conference, we will have

many more African decision-relevant applications presented