The carbon sink issues under the Kyoto Protocol had been negotiated intensively since COP3 until the agreement at COPT. Although, one of the largest carbon sink activities (3.9%) are allowed for Japan to achieve the reduction target (6%), there are still many issues need to be addressed regarding the implementation, such as additional forest management, development of the carbon accounting system etc. Especially, regarding the monitoring of the carbon sink activities, the Kyoto Protocol mandates each countries to report in a transparent and verifiable manner. In this regard, it is expected that the remote sensing techniques play a major role in establishing the carbon sink assessment system to meet the need of the Kyoto Protocol and beyond. This report review the international agreement associated with the carbon sink article, the cutting edge methods for the carbon sink assessment, and the related international research activities.

View full abstract

A method for estimating the canopy coverage of, forest by remote sensing techniques has been required. To illustrate, the Kyoto protocol adopted in 1997 has shown that it is important to calculate CO₂ absorption of the forest. To calculate CO₂, it is necessary to accurately estimate the canopy area of specific forest. Therefore, remote sensing techniques, which can observe wide ranges periodically, have been expected to estimate canopy coverage of specific forest.
In this paper, we propose an improved matched filter method for estimating canopy coverage of alder trees in a forest located in the Kushiro mire using remotely sensed imagery. The conventional matched filter method has been problem for estimating specific vegetation, because there are the effects of various vegetations within a single pixel. However, the canopy coverage of alder trees in the Kushiro was able to be estimated by the proposed method which removes the effects of other vegetation.

View full abstract

The Kyoto Protocol allows some carbon sequestration by vegetation sinks to be offset against CO₂ emissions. The sink projects under Joint Implementation (JI) and Clean Development Mechanism (CDM) are indispensable to achieve the emissions target of our country (-6% from the 1990 total).
In order to make the sink project successful, an evaluation of forestation and a monitoring and a verification of the carbon sink are critical factors. In this report, the impact study of forestation business in Green Triangle region of Australia was implemented by using multi-temporal Landsat/TM and ETM+ data. Green Triangle region is one of the potential areas of sink project for afforesttaion and reforestation of Article 3.3 clause of the Kyoto Protocol. The evaluation approach of the forest area based on the definition of the Kyoto Protocol was discussed for these areas.
Furthermore, we examined the practical use of satellite remote sensing technology for Kyoto Protocol by organizing the necessary condition of the sink and defining a process and items needed for monitoring.

View full abstract

Todevelop a monitoring system for carbon sinks under the Kyoto Protocol, a method of forest stand measurement using Airborne Lidar sensor is examined. A canopy DSM (Digital Surface Model) was constructed from first pulse of the lidar data, and the DTM (Digital Terrain Model) was constructed from the last pulse of the lidar data. The tree height surface was made from the difference of the canopy DSM and DTM. Crowns of trees were extracted from the tree height surface by the Watershed method. Numbers of stand trees were counted from the crowns. Tree heights were measured from height surface within the crown. Stem volumes of trees were calculated from the relationship between tree height and stem volume, from a yield-density effect curve, and from an aerial photographical analysis table. Carbon weights were calculated by using the relationship between the stem volumes and dry weights and the coefficient of carbon inclusion.
The results are : 1) The measurement error of tree height of Larix leptolepis stand was less than about 0.3 m (accuracy) and 1.0 m (precision). The precision of the measurement was assumed to be originated from the precision of truth data ; 2) the extracted crowns of trees were canopy trees, not understory trees, and the number of extracted crown had a tendency to be underestimated ; 3) stem volumes and carbon weights were underestimated by using the relationship between the tree heights and stem volumes and the yield-density effect curve. Using aerial photographical analysis table, however, yielded better results.
The next steps of this study will be 1) to develop of a method of the extraction technique from the lidar points with more accuracies, and 2) to develop of a method of aerial photographical analysis to estimate carbon weights for wide areas.

View full abstract

Integrated monitoring systems for forest biomass are necessary to estimate carbon stock changes in a forest area. This paper introduces airborne remote sensing technology to estimate the forest biomass and overview the development of aircraft platform for multiple sensors.
Although no specific forestration guidelines have been decided to utilize a cost-effective approach to ensure carbon dioxide emission curbing targets, technology development to detect carbon stock change should be carried out as priority strategy. This will bring the business opportunity, and also will contribute to global warming countermeasures.

View full abstract

In natural woody conditions, measuring tree position and diameter at breast height (DBH) of each Larix leptolepis by a portable scanning lidar is not easy because the tree is covered with not only their own thick branches but also other plants such as small broad-leaved trees, vines, epiphytes and quite tall ferns. Therefore, new methods for mapping tree position of Larix leptolepis woods and estimating the DBH using the range data of measurable tree parts obtained by a portable scanning lidar with high spatial resolution were examined. Furthermore, biomass of the trees was estimated from the DBH.
Using the range data measured by the portable scanning lidar that was set at a point in the woods, each tree position was mapped. Within the area of the angle of 170 degrees horizontally and 10m from the lidar, the number of measurable trees was 14 (100% of all trees in the area), within the area of 15m the number was 24 (83%), within the area of 20m the number was 29 (66%), and within the area of 30m the number was 44 (45%). Then the DBH of each tree was estimated in error of RMSE=7.3mm from the trunk diameter measured at a certain height of the tree using Eq. (1), where the coefficient k (h) had been obtained beforehand. Also, the biomass (fresh, dry and carbon weight) of each tree was estimated from the DBH using a quadratic equation (R²=0.96) correlated between DBH and biomass above ground. Furthermore, the biomass per square meter was estimated in error of 2.7% within the area of 30m from the lidar. The fresh weight and the carbon weight were obtained from the dry weight multiplied by 1.7 and 0.45 respectively.

View full abstract

One of the serious problems of the Kyoto Protocol is that we have no effective method to estimate the carbon stock of the subsurface. To solve this problem, we propose the application of ground-penetrating radar (GPR) to the subsurface soil survey. As a result, it is shown that GPR can detect the soil horizons, stones and roots. The fluctuations of the soil horizons in the forest are clealy indicated as the reflection pattern of the microwaves. Considering the fact that the physical, chemical, and biological characteristics of each soil layer is almost unique, GPR results can be used to estimate the carbon stock in soil by combining with the vertical soil sample survey at one site. Then as a trial, we demonstrate to estimate the carbon content fixed in soil layers based on the soil samples and GPR survey data. we also compare this result with the carbon stock for the flat horizon case. The advantages of GPR usage for this object are not only the reduction of uncertainty and the cost, but also the environmental friendliness of survey manner. Finally, we summarize the adaptabilities of various antennas having different predominant frequencies for the shallow subsurface zone.

View full abstract

In this article, our object is to predict biochemical contents (chlorophyll, water, nitrogen, lignin and cellulose) of larch needles from their spectral characteristics, because larch tree is a kind of important trees as sink for CO2. We firstly tested whether conifer leaf Model LIBERTY, which was constructed on jack pine and slash pine needles by Dawson et al., (1998), can predict quasi-infinite reflectance for larch needles from biochemical contents or not. And we established the way to predict respectively biochemical contents from global absorption coefficient, because LIBERTY Inversion model, which was constructed by Dawson et al., can predict just global absorption coefficient for cells from quasi-infinite reflectance. In conclusion, we could predict respectively biochemical contents from quasi-infinite reflectance for larch needles, and we suggested that there is the probability to predict them more accuracy, if 1) we modify LIBERTY model more suitable for larch needles, 2) incorporate carotenoid absorption coefficient into the model 3) get more data sets for larch needles and so on.

View full abstract

The hyper-spectral reflectance factor over the forest canopy is expected to obtain the information, which is related to the health condition, leaf biochemical contents ratio and photosynthetic activity. A hyper spectral digital camera system was developed in order to automatically take the hyper-spectral images of the forest in the daytime. This paper describes the outline of this camera system and initial result of acquired images of Tomakomai Flux research site. Spectral diffraction camera and autorotation stage and mainly constitute this system.
The camera has been installed in the top of the tower of the 25 m heights. The image acquisition is carried out from 8a.m. to 6p.m. in the 2-hour interval. This system covers the wavelength range from approximately 520 nm to 850 nm with a spectral resolution of 5nm. In 2001, the observation was carried out between September to the end of November.

View full abstract

A prototype fisheye camera is a hemispheric radiometer that was developed as a proto-flight model for the hot spot observation in canopy reflectance. This camera consists of the CCD camera and the fisheye lens that can measure the surface radiation with high resolution in instantaneous field of view in a short time. The averaged image calculated from a large number of the images measured in a short time shows detailed characteristics of surface reflectance in hot spot. This paper describes the design, calibration, and first measurements, which were carried out in visible and near infrared region at the Tomakomai flux research site, of the prototype fisheye camera.

View full abstract

Absorbed photosynthetically active radiation (APAR), which is defined as downward solar radiation in 400-700 nm absorbed by vegetation, is one of the significant variables for Net Primary Production (NPP) estimation from satellite data. Toward the reduction of the uncertainties in the global NPP estimation, it is necessary to clarify the APAR accuracy. In this paper, first we proposed the improved PAR estimation method based on Eck and Dye's method in which the ultraviolet (UV) reflectivity data derived from Total Ozone Mapping Spectrometer (TOMS) at the top of atmosphere were used for clouds transmittance estimation. The proposed method considered the variable effects of land surface UV reflectivity on the satellite-observed UV data. Monthly mean PAR comparisons between satellite-derived and ground-based data at various meteorological stations in Japan indicated that the improved PAR estimation method reduced the bias errors in the summer season. Assuming the relative error of the fraction of PAR (FPAR) derived from Moderate Resolution Imaging Spectroradiometer (MODIS) to be 10%, we estimated APAR relative errors to be 10-15%. Annual NPP is calculated using APAR derived from MODIS/ FPAR and the improved PAR estimation method. It is shown that random and bias errors of annual NPP in a 1 km resolution pixel are less than 4% and 6% respectively. The APAR bias errors due to the PAR bias errors also affect the estimated total NPP. We estimated the most probable total annual NPP in Japan by subtracting the bias PAR errors. It amounts about 248 MtC/yr. Using the improved PAR estimation method, and Eck and Dye's method, total annual NPP is 4% and 9% difference from most probable value respectively. The previous intercomparison study among using fifteen NPP models⁴⁾ showed that global NPP estimations among NPP models are 44.4-66.3 GtC/yr (coefficient of variation=14%). Hence we conclude that the NPP estimation uncertainty due to APAR estimation error is small compared with model uncertainties.

View full abstract

Past, present and future carbon cycle was analyzed based on earth observation satellite data and a simplified carbon cycle-climate coupled model. A simple Earth system model, the Four Spheres Cycles of Energy and Mass model (4-SCEM) was developed to simulate global warming due to anthropogenic CO₂ emission. The model consists of the Atmosphere-Earth Heat Cycle model (AEHC), the Four Spheres Carbon Cycle mode (4-SCC), and their feedback processes. The following feedback processes were included in the model, (1) water vapor feedback, (2) biospheric CO₂ fertilization, and temperature dependencies on (3) photosynthesis, (4) soil decomposition, and (5) ocean surface chemistry. The recent NPP trends derived from the 4-SCEM and satellite data were directly compared for validation. The satellite-based estimation showed the NPP increase in past 20 years at a rate of 3.8% per 10 years. Although the 4-SCEM based analysis also showed a recent increase in NPP, the result was a half (with climate feedback) or one-eighth (without climate feedback) of satellite based one. Although large discrepancies are still remained, we can conclude that the carbon cycle model with climate feedback is more reasonable than that without climate feedback. The future global carbon cycle and climate was simulated under the IS92a emission scenario. The atmospheric CO₂ concentration reaches 712 ppmv in 2100 and a peak of 910 ppmv, and becomes a steady state with 650 ppmv. Although the climate feedback effects on the future atmospheric CO₂ were not large in 4-SCEM, high sensitivity was found in the terrestrial NPP. If we include the nutrient limitation effect in terrestrial carbon cycle model, the climate feedback effect on atmospheric CO₂ may be amplified due to NPP saturation.

View full abstract