A framework for creating and validating a non-linear spectrum-biomass model to estimate the secondary succession biomass in moist tropical forests

Li, Hui; Mausel, Paul W.; Brondı́zio, Eduardo; Deardorff, David

doi:10.1016/j.isprsjprs.2010.01.002

Cited by 22 publications

(14 citation statements)

References 53 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…This assumption is usually not met in practice because remotely sensed data are often highly correlated each other [16]. There is also an issue with the regression approach in that the selected variables may have a nonlinear relationship with forest biomass [26]. An alternative to these limitations is to use nonparametric approaches such as neural network and KNN.…”

Section: Biomass Estimation With Optical Sensor Datamentioning

confidence: 99%

Aboveground Forest Biomass Estimation with Landsat and LiDAR Data and Uncertainty Analysis of the Estimates

Chen

Wang

et al. 2012

International Journal of Forestry Research

196

156

View full text Add to dashboard Cite

Landsat Thematic mapper (TM) image has long been the dominate data source, and recently LiDAR has offered an important new structural data stream for forest biomass estimations. On the other hand, forest biomass uncertainty analysis research has only recently obtained sufficient attention due to the difficulty in collecting reference data. This paper provides a brief overview of current forest biomass estimation methods using both TM and LiDAR data. A case study is then presented that demonstrates the forest biomass estimation methods and uncertainty analysis. Results indicate that Landsat TM data can provide adequate biomass estimates for secondary succession but are not suitable for mature forest biomass estimates due to data saturation problems. LiDAR can overcome TM's shortcoming providing better biomass estimation performance but has not been extensively applied in practice due to data availability constraints. The uncertainty analysis indicates that various sources affect the performance of forest biomass/carbon estimation. With that said, the clear dominate sources of uncertainty are the variation of input sample plot data and data saturation problem related to optical sensors. A possible solution to increasing the confidence in forest biomass estimates is to integrate the strengths of multisensor data.

show abstract

Section: Biomass Estimation With Optical Sensor Datamentioning

confidence: 99%

Aboveground Forest Biomass Estimation with Landsat and LiDAR Data and Uncertainty Analysis of the Estimates

Chen

Wang

et al. 2012

International Journal of Forestry Research

196

156

View full text Add to dashboard Cite

show abstract

“…After the regression the coefficient of determination, or called Rsquared (R 2 ), is calculated to evaluate how well the data fits the statistical model. The adjusted R 2 is applied to obtain an unbiased estimation of R 2 by taking into account the degree of freedom (Li et al 2010). …”

Section: Lai L Pimentioning

confidence: 99%

“…Although remote sensing techniques are highly suitable for a wide range of observations (Lillesand et al 2004;Li et al 2010;Kumar et al 2013), mapping forest structure has been challenging for early remote sensing systems with existing optical and radar sensors due to the limitations of sensing vertical information. Optical and radar systems can provide canopy surface height information based on multi-angle data, but observations for the forest canopy interior are frequently invalid.…”

Section: Introductionmentioning

confidence: 99%

Mapping CHM and LAI for Heterogeneous Forests Using Airborne Full-Waveform LiDAR Data

Tseng¹,

Lin²,

Wang³

2016

Terr. Atmos. Ocean. Sci.

View full text Add to dashboard Cite

Canopy height model (CHM) and leave area index (LAI) are essential forest structure attributes that are estimated to understand the ecological states and processes occurring in forest ecosystems. Airborne light detection and ranging (LiDAR) systems have proven efficient in producing both CHM and LAI maps for heterogeneous forests at the regional scale. The unique advantage of airborne LiDAR over optical and radar sensors is its vegetation penetration capability. Although the LiDAR penetration capability decreases in dense, complex forests, full-waveform LiDAR systems are currently available to provide critical point observations under the forest canopy. This research developed and tested methods to map CHM and LAI in heterogeneous forests using airborne waveform LiDAR datasets acquired using two different LiDAR systems and flight altitudes. Since using waveform data significantly increases the laser penetration rate, the test results strongly recommend using waveform data for the estimation of both CHM and LAI. These experiments also revealed that the flight data collection altitude will not affect LAI estimation. Through the analysis of CHMs and LAI data variations derived from 4 different datasets, CHM estimation may be good to 0.8 m and LAI estimation may be as precise as 0.5.

show abstract

“…In this assessment, we adopted a simple nonlinear regression model to conduct a preliminary assessment of the relationship between the above ground biomass and selected reflectance values (Hall et al, 2006;Popescu, 2007;Zheng et al, 2004). The decision was based on the examination of the scatter plot and correlation coefficient of all the spectral bands and major vegetation indices, which elucidated a mainly curved or nonlinear relationship between the variables of interest (Li et al, 2010). The method selected will address the issue of nonlinear relationships between spectral responses or vegetation indices with estimated biomass.…”

Section: Aboveground Biomass Estimationmentioning

confidence: 99%

“…Therefore the error and accuracy assessment of the application was based on the evaluation of the coefficient of determination values (R 2 ) and the predictive capability of the developed regression model (Popescu, 2007). It was performed by comparing differences in R 2 and root mean square error (RMSE) (Li et al, 2010).…”

Section: Aboveground Biomass Estimationmentioning

confidence: 99%

Integrating a REDD+ Project into the management of a production mangrove forest in Matang Forest Reserve, Malaysia

Aziz¹

View full text Add to dashboard Cite

Whilst much attention has recently been given in the policy and scientific literature to the theoretical opportunities that REDD+ and blue carbon initiatives can potentially provide for sustainable mangrove forest management, little attention has been paid to the practicalities of how carbon payments for conservation can be best balanced with other somewhat conflicting forest use activities, such as wood production. This thesis explores these gaps in the literature by analysing the integration of a hypothetical REDD+ pilot project into the management planning of the Matang Mangrove Forest Reserve in Malaysia.The Matang Mangrove Forest reserve covers 40,466 ha and has been managed for more than a century primarily for the extraction of wood for the production of poles and charcoal. The forest reserve has often been referred to as a fine example of sustainable forest management. In recent years, evidence has emerged of the decline of multiple ecosystem services in the Matang forest reserve area. The premise to this thesis is that something should be done to arrest the decline of ecosystem health in the reserve. The thesis explores whether carbon payments can be used to protect areas within the reserve from harvesting for wood production and thereby support the recovery and improved health of the ecosystem, and at the same time, support income for the forest owners and livelihood activities (particularly through wood production) for local communities.The thesis begins by reviewing the potential benefits and common problems associated with integrating a REDD+ pilot project in a mangrove forest such as the Matang Mangrove ForestReserve. The suitability of remotely-sensed data obtained from Landsat imaging sensors, in combination with ancillary forest inventory data, is then assessed for planning and management in a production mangrove forest. Different techniques for the spatial analysis of the ecological features of a mangrove forest are tested, evaluated and presented. The techniques include both pixel-based and object-based approaches. The object-based approach accurately delineated between commercial and non-commercial mangrove species. More importantly, it also produced a reliable and accurate stand-age structure map based on three designated classes: clear-felled areas; young mangrove forest (less than 13 years old); and mature mangrove forest (13 years or older).The technique was then applied to assess the impact of the silvicultural practices employed in the Matang forest reserve on the provision of the critical ecosystem services. The spatiotemporal changes in the distribution of the stand age composition and structure over a 35 year period was iii analysed. Results revealed a relatively stable composition of clear felled areas, young and mature mangrove forests throughout the observation period -an indication of good area and composition governance by the Matang reserve managers (the State Forestry Department of Perak). However, the results of the ecosystem services assessment revealed that the rese...

show abstract

A framework for creating and validating a non-linear spectrum-biomass model to estimate the secondary succession biomass in moist tropical forests

Cited by 22 publications

References 53 publications

Aboveground Forest Biomass Estimation with Landsat and LiDAR Data and Uncertainty Analysis of the Estimates

Aboveground Forest Biomass Estimation with Landsat and LiDAR Data and Uncertainty Analysis of the Estimates

Mapping CHM and LAI for Heterogeneous Forests Using Airborne Full-Waveform LiDAR Data

Integrating a REDD+ Project into the management of a production mangrove forest in Matang Forest Reserve, Malaysia

Contact Info

Product

Resources

About