Uncertainty [MGD Sections]

UNFCCC decisions and requirements
IPCC good practice guidance
Relationship to UNFCCC
GHGI coverage, approaches, methods and tiers
Design decisions relevant to national forest monitoring systems
Land cover, land use and stratification
Forest reference emission levels and forest reference levels
Quality assurance and quality control
Guiding principles – Requirements and design decisions
Estimation methods for REDD+ activities
Integration frameworks for estimating emission and removals
Selecting an integration framework
Activity data x emission/removal factor tools
Fully integrated tools
Practical considerations in choosing an integration tool
Guiding principles – Methods and approaches
Remote sensing observations
Coarse resolution optical data
Medium resolution optical data
High resolution optical data
L-band Synthetic aperture radar
C-band and X-band SAR
LIDAR
Global forest cover change datasets
Ground-based observations
National forest inventories
Auxiliary data
Guiding principles – Remote sensing and ground-based observations
Activity data
Methods for estimating activity data
Maps of forest/non-forest, land use, or forest stratification
Detecting areas of change
Additional map products from remote sensing
Estimating uncertainty of area and change in area
Estimating total emissions/removals and its uncertainty
REDD+ requirements and procedures
Reporting forest reference emission levels and forest reference levels
Technical assessment of forest reference emission levels and forest reference levels
Reporting results of REDD+ activities
Technical analysis of the REDD+ annex to the BUR
Additional advice on REDD+ reporting and verification
Guiding Principles – Reporting and verification of emissions and removals
Financial considerations
Country examples – Tier 3 integration
Use of global forest change map data
Relative efficiencies
Developing and using allometric models to estimate biomass

Record Keeping [MGD Sections]

Integration + Estimation [MGD Sections]

Ground Based Observations [MGD Sections]

4.1.6   LIDAR Previous topic Parent topic Child topic Next topic

LIDAR sensors emit pulses in near-infrared wavelengths that interact with different strata and from which quantitative information on forest structure (e.g. tree height, canopy volume) and biomass can be estimated. LIDAR-assisted biomass estimation using wall-to-wall coverage of satellite data is a research topic of interest for future forest monitoring systems. Although an historic archive of satellite LIDAR is available(1), there are currently no operational LIDAR satellites. The NASA ICESAT-2 mission for a space-borne LIDAR is planned to be launched in 2017, and the GEDI mission is scheduled for launch before 2020. Subject to the demonstration of suitable techniques, space-borne LIDAR could be used for estimation and cross-checking with other methods. Airborne LIDAR, following calibration using appropriate ground estimates of biomass, can be used to produce reliable high resolution biomass maps (Stephens et al., 2011(2), Goslee et al., 2015 Opens in new window) and can be cost effective in some national circumstances, e.g. where terrain makes access difficult. Box 16 and Box 17 summarize practical experiences with LIDAR, in Nepal and Tanzania respectively.

Box 16: Use of LIDAR by Nepal

The Department of Forest Resource and Survey (DFRS) of Nepala conducted a comparison between the use of airborne LIDAR, high resolution RapidEye satellite data versus conventional ground-based techniques for estimating above-ground biomass.
The first approach applied a model-based LIDAR Assisted Multisource Program (LAMP), integration 5% LIDAR sampling, wall to wall Rapid Eye satellite images, and in situ measurements from 738 field sample plots of 12.62 m radius (in LIDAR sample areas) over the 23300 km2 Terai Arc Landscape (TAL) area of Nepal during March to May 2011 to estimate AGB.
In the second approach the field based multisource forest resource assessment (FRA) began in January 2011. The field based approach utilised a design-based forest monitoring method incorporating space technology, auxiliary data, and intensive field inventory. A total of 676 concentric circular plots (CCP) of radii 20m, 15m, 8m and 4m were designated systematically in TAL area to measure tree characteristics, including the attributes required for calculating AGB.
Both field plot-based FRA method and the LIDAR assisted LAMP approach were compared with respect to their accuracy in estimating mean AGB for the region at different spatial scales. The mean error of the FRA estimated at 1 ha was 6243.95 tons/ha which is impossibly high, but this decreases slowly with increasing estimation area and goes down to 10.6 tons/ha when the estimation area reaches 350,000 ha. The mean error for the LAMP approach was 13.21 tons/ha for 100 ha of forest which demonstrates acceptable accuracy to estimate biomass stock in management level forest regime such as community managed forests of the TAL area where the average size of community forests is 150 ha. Error calculation for the two approaches shows the importance of considering national circumstances (in this case accessibility and typical size of community forests) in deriving national approaches (Kandel et al., 2013).
Whilst the LAMP approach achieved lower uncertainty, the FRA approach had lower baseline data collection costs. However, LAMP was found to be the more cost effective approach for repeated forest monitoring required for MRV.
The results show that the biggest difference between the two approaches is spatial resolution. LAMP has higher accuracy reliability over smaller spatial extent compared to conventional multisource forest inventory.
This study reinforced that choice of inventory method should be made depending on the reason for the inventory (e.g. MRV vs. forest industry management) and the cost of measuring forest variables. Through the FRA method, information about a vast number of target variables can be collected, ranging from tree-level characteristics to biodiversity and soil. The LAMP method covers significantly fewer forest variables and cannot replace a multisource inventory. However, LAMP produces biomass and carbon stock estimates at high spatial resolution. For estimation of forest biomass/ carbon stock and establishing an MRV baseline, LIDAR-assisted inventory was preferred because subsequent monitoring cost is low.
Notes: a. in collaboration with WWF-Nepal and Arbonaut

Box 17: Use of LiDAR in Tanzania

Miombo woodlands are the dominant forest type in eastern Africa and occupy around 9% of the African land surface south of Sahara. They account for approximately 90% of the land area of Tanzania. Miombo woodlands are fragmented with mainly open forests, and in many areas they are subject to severe degradation and conversion to agricultural land.
Tanzania has recently established an NFI consisting of over 30,000 ground plots across all land use categories distributed according to a stratified and systematic sampling scheme on a 5-km × 5-km grid. To explore the possibilities of enhancing the precision of AGB and hence carbon stock estimates, and stock change estimates, airborne scanning LiDAR was used as a sampling device in combination with the existing NFI. Previous studies had shown that it can be efficient to sample with LiDAR as opposed to collecting wall-to-wall data due to reduced costs and only marginal reduction of the precision of estimates.
For trial purposes a 16,000 km2 miombo woodland area in Liwale district of SE Tanzania, was subject to LiDAR strip sampling. Thirty-two parallel LiDAR strips were distributed in the E-W direction with a distance between strips of 5 km. Each of the LiDAR strips was wider than 1 km and covered >25% of the total area. LiDAR data were collected in 2012 and repeat LiDAR measurements were acquired along the same strips in 2014. Coincident ground observations were obtained on 531 NFI plots along the strips in 2012 as well as in 2014.
Numerous analytical estimators for biomass and variance of biomass estimates for different types of LiDAR-based sampling applications have been derived over the past 10 years. Some assume probability sampling of LiDAR strips and field data, while others assume probabilistic sampling of LiDAR data and permit opportunistic collection of ground data in areas that are accessible. Different types of estimators were used to estimate AGB and change in AGB over a period of two years.
It was shown that LiDAR could greatly improve the precision of AGB estimates. The LiDAR-assisted estimated mean AGB across the entire area was 59.7 Mg ha-1 with a standard error (precision) of 1.73 Mg ha-1 for the LiDAR-assisted estimate and 4.79 Mg ha-1 when using only the field data. It was shown that by using LiDAR in combination with the field survey, the overall costs could be reduced while maintaining precision by reducing the field sampling effort. However, given that the NFI is a continuous inventory system, reducing field efforts may not be an option. Also, an NFI provides information on many other aspects of forests than just biomass which cannot be acquired by remote sensing.
Change estimates over the 2-yr period showed a loss of biomass of 0.26 Mg ha-1 (0.22% per year). The standard error was 0.81 Mg ha-1. An approximate 95% confidence interval would range from -1.36 to 1.88 Mg ha-1, i.e., spanning an interval from increase in stock to a decrease in stock. As opposed to the stock estimates, the precision of the change estimates did not improve by using LiDAR in addition to the field plots. Consequently, the use of LiDAR for change estimation was not cost-efficient. More information can be found in Forest monitoring with airborne laser scanning in Tanzania, Erik Næsset (editor), INA fagrapport 31, ISSN, Norwegian University of Life Sciences 2015 (ISSN 1891-2281).

 (1)
 (2)
When building the national reporting system for New Zealand concluded that airborne LiDAR with appropriate plots in a bound-sample found that a "reduction in standard error achievable using LiDAR is expected to be approximately 50% for total carbon and 55% for AGB carbon.