use of lidar for estimating reference emission level in nepal s.k. gautam dfrs, nepal
TRANSCRIPT
Use of Lidar for estimating Reference
Emission Level in Nepal
S.K. Gautam DFRS, Nepal
Nepal’s ER-Program covers 12 jurisdictional Terai districts out of 75 districts of the country;
Total area under the ER-Program is 2.3 million ha (about 15% of the country);
About 52% of the ER-Program area (1.18 million ha) is under different types of forest cover.
The area is linked with eleven trans-boundary protected areas across Nepal and India and is home to flagship species like tigers, rhinos, Asiatic wild elephants, and many other endangered species.
Total population of the ER-Program area is 7.35 million and constitute about 27% of total population (2011 population census)
2Introduction: Study Area
Samples (5%) of LiDAR data to calibrate satellite models;
Reference field sample plots to calibrate/validate LiDAR models;
Landsat satellite imagery for wall-to-wall biomass map.
3Introduction: LAMP
Stratification from a Landsat-based forest classification map.
Weight calculated for every block as a product of the importance of the forest types and the inverse of the forest types area.
The forest classification was used as a priori information to calculate weighting function for random block and systematic plot design.
5 km x 10 km systematic grid over the study area
where ew is the expert weight and A is the area
LiDAR block design
Forest type Area, km2 Expert weight Area-normalized weightHill-sal 3625 100 541Sal 3458 200 1135Mixed 1299 200 3020Riverine 180 100 10880Grass 873 50 1124Degraded forest 1098 50 893Chir pine 442 100 4436Shadow 598 100 3283Non-forest 8043 0 0
Forest type map with forest type weights. The larger weights are withbrighter tones in gray-scale. Black = zero weight (non-forest).
LiDAR block design
The basic math is: • Activity Data (ha change/year) × Emissions Factors
(tCO2/ha)
= tCO2/year
• Activity data is be based on satellite information (past) or assumptions (future)
• Emissions factors are based on field measurements and allometric equations
Basics of a REDD+ RL
• Defines forest/non-forest for 1999 inception date of RL with 1998 Topographic basemaps
• Utilizes satellite analysis for 1999, 2001, 2006, 2009 and 2011 to delineate structural classes of intact, degraded and deforested
• Bases classification on fractional image indexes (i.e., % vegetation) and temporal analysis drawing on work by leader in the field Carlos Souza of Brazil
• Develops land cover change matrix by tracking changes between the different structural classes between 4 time-periods
Activity Data
• DFRS, FRA, Arbonaut and WWF collaborate in collection of LiDAR data covering 5% of TAL program area in 2011
• Field plots collected in 2011 (738 calibration plots) and 2013 (46 validation plots)
• Uses allometric equations of Sharma and Pukkala (1990) to estimate biomass for ground plots (same equations used by FRA)
Emissions Factors
• Model to correlate LiDAR-based above-ground biomass estimates for each forest condition (intact, deforested, degraded and regeneration) and forest type (Sal, Sal mixed, Other mixed and Riverine)
• IPCC default values used to calculate mean carbon density for regeneration and below-ground carbon based on biomass estimates
Emissions Factors
Method: LAMP Workflow
CO2 Emissions
(tCO2e)Period Above-
groundBelow-ground
Total
1999-2002 13,136,430 2,627,286 15,763,716
2002-2006 1,736,537 347,307 2,083,845
2006-2009 9,644,698 1,928,940 11,573,637
2009-2011 19,020,661 3,804,132 22,824,793
Total 12-yr
43,538,325
8,707,665
52,245,991
Average annual
3,628,193
725,639 4,353,833
Average annual net CO2 emissions (tCO2e) in TAL between 1999 and 2011.
Results: Historical CO2 emissions
19992001
20032005
20072009
2011 85,000,000
90,000,000
95,000,000
100,000,000
105,000,000
110,000,000 to
ns o
f C
Results: Historical Carbon Stock Loss
a) Comparison to independent field plots
b) Leave-one-out validation
Accuracy assessment
LiDAR model LAMP model (Landsat)
Field-measured biomass (t/ha)
Est
imate
d b
iom
ass
(t/
ha)
R2 = 0.92 R2 = 0.52
Emissions reduction targets
Intervention
Cumulative emissions reduction from BAU (millions of tons CO2e)
5 years (2015 -2020)
10 years(2015 -2025)
15 years(2015 -2030)
Sustainable management of forests 9.9 29.2 49.0
Installed biogas plants 0.9 3.4 6.5
Improved cook stoves 0.3 1.1 2.0
Land use planning 2.8 8.3 13.9
Private forestry/tree nurseries 0.1 0.7 1.4
Total 14.0 42.7 72.8
Endorsed by the FCPF
Research & Development: AGB in 2010
Research & Development: AGB in 1999
Research and Development: Difference in AGB between 1999-2011
The cost of this project is USD 0.28/ha
Our experience shows that 1-2% LiDAR coverage is sufficient for this integrated approach
But LiDAR is needed only once
Subsequent monitoring is based on new satellite images to which the LAMP models are applied
Costs and Future Monitoring
Basic image processing steps in ImgTools
Decision Tree and Definition of Forest for Terai Arc Landscape
Four major forest types: 1) Sal forest, 2) Sal dominated mixed forest, 3) other than Sal dominated forest (i.e. “other forest”) and 4) Riverine.
The four forest types were overlaid on the forest structural map (Joshi et al. 2003) to generate forest types and conditions maps for each time period.
The study assumed forest types do not change from one type to another type (i.e., from Sal forest to mixed forest or riverine forest or vice versa) in 10-20 years;
Forest types and conditions map
Thank You