Final Report Summary - GESAPU (Geoinformation technologies, spatio-temporal approaches, and full carbon account for improving accuracy of GHG inventories)
Greenhouse gases (GHG) inventories involve uncertainties, due to the fact that (1) some data are unavailable, imprecise, and/or incomplete and (2) limited knowledge of a number of important processes, particularly in the LULUCF sector. These uncertainties might be ineluctable, and many of them are significant and have important scientific, and policy implications: (1) The compliance with commitments to reduce GHG emissions is unsolved, although for most countries the emission changes agreed on under the Kyoto Protocol are often smaller than the uncertainty in their estimated emissions; (2) None of the emission trading systems deals with the uncertainty of emissions estimates; (3) Existing methodologies of carbon cycling account in the LULUCF sector, as a rule, do not allowing the estimation of structural uncertainties, i.e. assessing of “uncertainties of uncertainties”.
The schemes introduced by the Kyoto Protocol have also a number of other principal gaps that substantially hinder the chance to reach the goals of the UN FCCC, like: (1) a distortion of the real picture of the role of individual countries in climate change mitigation efforts, because a substantial part of emissions and removals of greenhouse gases are not included in the accounting regime; for large regions of the Earth the omitted part may provide emissions that exceed those from industry and “managed” part of the biosphere; (2) a threat to the protection of some categories of “unmanaged” ecosystems, e.g. old growth forests; (3) an unsatisfactory consideration of large sources of emissions (e.g. wild fires). More precisely, partial accounts do not allow any comprehensive analysis of uncertainties due to the fact that, considering their impacts on part of a system, they are not sufficient for assessing the responses and feedbacks of the entire system in any complete form.
Overall, decreasing inventories’ uncertainty is of paramount importance to its credibility, quality of compliance, and proper functioning of trading systems. Quantification of CO2 emissions at fine spatial scales is also advantageous for many environmental, physical, and socio-economic analyzes; in principle it can be easily integrated with other data in gridded format. This project aim is to improve accuracy of the inventories at the national level, particularly in countries with differentiated regions, and/or substantial areas of forests i.e. in the case when standard inventory procedures fail in producing high accuracy assessments. The proposed solution is to apply spatially distributed inventories and more solid and detailed forest modeling.
One of the objectives of the project is to develop a spatial inventory of technogenic GHG emissions for Poland. However, crucial activity data of a required spatial resolution are unavailable in Poland. The proposed method relies on statistical downscaling models, which are novel in providing GHG inventories. The methodology uses multilevel mathematical models, digital maps, geo-referenced databases, and appropriate software.
GIS rooted georeferenced databases has been created, which allows for creating spatial emission inventories of GHGs in Poland from the IPCC Guidelines sectors and subsectors, as well as total emissions, with the very high resolution. In the project inventories in 2010 with the resolution 2 km × 2 km have been compiled. The emissions from the energy sector have been compared with an independent estimates created in the ODIAC project with the resolution around 1 km × 1 km using nighttime lights. Good match has been obtained except few percent of cells, where higher differences were caused by wrong allocation of high emission points in a CARMA database used in ODIAC project or due to ill-aligned grids in the compared maps.
Two novel methods for improved disaggregation of spatially distributed variables have been developed. One of them can be used for spatially correlated emissions, which is often the case of areal emissions. Taking into account the correlation, even few dozen percent improvement in comparison with the usually applied regression method can be achieved. The method can be used both for regular and irregular grids. Another method deals with combining information from ill-aligned maps. It uses completely different approach than usual area weighting or regression methods and relies on intelligent computation approach and fuzzy logic reasoning. This way different “soft”, non-numeric knowledge can be effectively employed in better approximate the emission distribution.
A review of possible sources of additional independent information, which can be used in improving emission estimates has been done. This information comes first of all from different measurements performed in observational sites. A Bayes approach to incorporating this information has been adopted and an idea of improving this additional information in reducing uncertainty of estimates has been outlined.
Further possibility of reduction of uncertainty was examined by modification and adaptation of the methodology, accounting schemes, and models of the verified full carbon account (FCA) of forest ecosystems. In essence, the FCA is a large open dynamic fuzzy system that comprises a sophisticated interplay of many stochastic elements/processes. Such systems cannot be directly validated or verified in any formal way due to evident labor and resource limitations. An appropriate methodology is based on system integration of methods and models of a different nature, harmonizing and multiple constraints of intermediate and final results received by independent methods. The landscape-ecosystem approach serves as a system background of the methodology. Forests are considered as an informative case study due to their highly complicated and poorly understood role in global carbon cycle as well as a complicated structure of forest ecosystems.
Forests in Ukraine play an extremely important role: (1) as a protective component of the environment (more than 50% of Ukrainian forests are represented by strictly protective areas); and (2) as a crucial stabilizing element of agroforestry landscapes. Within this context a verified FCA for Ukrainian conditions requires that uncertainties are defined reliably and comprehensively, and are below a preliminary defined level. Thus, the work aimed at developing advanced approaches to evaluate major components of the forest FCA and, consequently, to assess the biospheric role of forests in Ukraine as a typical country with the low forest cover in its substantial part.
Hybrid forest map of Ukraine was created with spatial resolution of 300 m. The map has been parameterized with main forest parameters (tree species, age, site index etc.) by using a method that integrates statistics, remote sensing and in-situ information. The accuracy of forest map was validated using (1) in-situ data of forest parameters obtained from forest inventory and (2) remote sensing crowd-sourcing data collected in geo-wiki environment (http://www.geo-wiki.org).
This study supported possibility of the FCA for Ukrainian forests based on a consistent use of major principles of applied systems analysis, integration of all relevant data in form of an Integrated Land Information System and use of advanced modeling. Major carbon pools (live biomass, coarse woody debris) and fluxes (heterotrophic soil respiration, fluxes caused by fire and other natural and human-induced disturbances, fluxes to hydrosphere and lithosphere) were estimated. Received uncertainty of the result (the net carbon sink at 11.0±2.4 Tg C year-1 over 1996-2011) seems acceptable for policy makers. However, there are evident limitations for further improving the uncertainties that depends on (1) the gaps of some important information, (2) the lack of results received by other methods of the FCA (e.g. eddy covariance, regional process-based models etc.), and (3) inevitable use of some expert estimates under the current level of knowledge.
The study considered linkages between sustainable forest/land use management practices and carbon management. Besides of the meaningful biospheric role, the Ukrainian forests play the extremely important role in protection of soil and water, particularly in agro-landscapes. Also the use of forest biomass for energy production is a substantial part of the problem that should be solved in Ukraine. Currently disturbances (wildfire, outbreaks of pests and diseases, illegal logging) substantially impact the carbon cycling of Ukrainian forests. Expected climate change might generate much more dangerous threat, particularly for forests which grow in the xeric belt. The study outlined major direction of adaptation to and mitigation of, the negative consequences of climate change.
The schemes introduced by the Kyoto Protocol have also a number of other principal gaps that substantially hinder the chance to reach the goals of the UN FCCC, like: (1) a distortion of the real picture of the role of individual countries in climate change mitigation efforts, because a substantial part of emissions and removals of greenhouse gases are not included in the accounting regime; for large regions of the Earth the omitted part may provide emissions that exceed those from industry and “managed” part of the biosphere; (2) a threat to the protection of some categories of “unmanaged” ecosystems, e.g. old growth forests; (3) an unsatisfactory consideration of large sources of emissions (e.g. wild fires). More precisely, partial accounts do not allow any comprehensive analysis of uncertainties due to the fact that, considering their impacts on part of a system, they are not sufficient for assessing the responses and feedbacks of the entire system in any complete form.
Overall, decreasing inventories’ uncertainty is of paramount importance to its credibility, quality of compliance, and proper functioning of trading systems. Quantification of CO2 emissions at fine spatial scales is also advantageous for many environmental, physical, and socio-economic analyzes; in principle it can be easily integrated with other data in gridded format. This project aim is to improve accuracy of the inventories at the national level, particularly in countries with differentiated regions, and/or substantial areas of forests i.e. in the case when standard inventory procedures fail in producing high accuracy assessments. The proposed solution is to apply spatially distributed inventories and more solid and detailed forest modeling.
One of the objectives of the project is to develop a spatial inventory of technogenic GHG emissions for Poland. However, crucial activity data of a required spatial resolution are unavailable in Poland. The proposed method relies on statistical downscaling models, which are novel in providing GHG inventories. The methodology uses multilevel mathematical models, digital maps, geo-referenced databases, and appropriate software.
GIS rooted georeferenced databases has been created, which allows for creating spatial emission inventories of GHGs in Poland from the IPCC Guidelines sectors and subsectors, as well as total emissions, with the very high resolution. In the project inventories in 2010 with the resolution 2 km × 2 km have been compiled. The emissions from the energy sector have been compared with an independent estimates created in the ODIAC project with the resolution around 1 km × 1 km using nighttime lights. Good match has been obtained except few percent of cells, where higher differences were caused by wrong allocation of high emission points in a CARMA database used in ODIAC project or due to ill-aligned grids in the compared maps.
Two novel methods for improved disaggregation of spatially distributed variables have been developed. One of them can be used for spatially correlated emissions, which is often the case of areal emissions. Taking into account the correlation, even few dozen percent improvement in comparison with the usually applied regression method can be achieved. The method can be used both for regular and irregular grids. Another method deals with combining information from ill-aligned maps. It uses completely different approach than usual area weighting or regression methods and relies on intelligent computation approach and fuzzy logic reasoning. This way different “soft”, non-numeric knowledge can be effectively employed in better approximate the emission distribution.
A review of possible sources of additional independent information, which can be used in improving emission estimates has been done. This information comes first of all from different measurements performed in observational sites. A Bayes approach to incorporating this information has been adopted and an idea of improving this additional information in reducing uncertainty of estimates has been outlined.
Further possibility of reduction of uncertainty was examined by modification and adaptation of the methodology, accounting schemes, and models of the verified full carbon account (FCA) of forest ecosystems. In essence, the FCA is a large open dynamic fuzzy system that comprises a sophisticated interplay of many stochastic elements/processes. Such systems cannot be directly validated or verified in any formal way due to evident labor and resource limitations. An appropriate methodology is based on system integration of methods and models of a different nature, harmonizing and multiple constraints of intermediate and final results received by independent methods. The landscape-ecosystem approach serves as a system background of the methodology. Forests are considered as an informative case study due to their highly complicated and poorly understood role in global carbon cycle as well as a complicated structure of forest ecosystems.
Forests in Ukraine play an extremely important role: (1) as a protective component of the environment (more than 50% of Ukrainian forests are represented by strictly protective areas); and (2) as a crucial stabilizing element of agroforestry landscapes. Within this context a verified FCA for Ukrainian conditions requires that uncertainties are defined reliably and comprehensively, and are below a preliminary defined level. Thus, the work aimed at developing advanced approaches to evaluate major components of the forest FCA and, consequently, to assess the biospheric role of forests in Ukraine as a typical country with the low forest cover in its substantial part.
Hybrid forest map of Ukraine was created with spatial resolution of 300 m. The map has been parameterized with main forest parameters (tree species, age, site index etc.) by using a method that integrates statistics, remote sensing and in-situ information. The accuracy of forest map was validated using (1) in-situ data of forest parameters obtained from forest inventory and (2) remote sensing crowd-sourcing data collected in geo-wiki environment (http://www.geo-wiki.org).
This study supported possibility of the FCA for Ukrainian forests based on a consistent use of major principles of applied systems analysis, integration of all relevant data in form of an Integrated Land Information System and use of advanced modeling. Major carbon pools (live biomass, coarse woody debris) and fluxes (heterotrophic soil respiration, fluxes caused by fire and other natural and human-induced disturbances, fluxes to hydrosphere and lithosphere) were estimated. Received uncertainty of the result (the net carbon sink at 11.0±2.4 Tg C year-1 over 1996-2011) seems acceptable for policy makers. However, there are evident limitations for further improving the uncertainties that depends on (1) the gaps of some important information, (2) the lack of results received by other methods of the FCA (e.g. eddy covariance, regional process-based models etc.), and (3) inevitable use of some expert estimates under the current level of knowledge.
The study considered linkages between sustainable forest/land use management practices and carbon management. Besides of the meaningful biospheric role, the Ukrainian forests play the extremely important role in protection of soil and water, particularly in agro-landscapes. Also the use of forest biomass for energy production is a substantial part of the problem that should be solved in Ukraine. Currently disturbances (wildfire, outbreaks of pests and diseases, illegal logging) substantially impact the carbon cycling of Ukrainian forests. Expected climate change might generate much more dangerous threat, particularly for forests which grow in the xeric belt. The study outlined major direction of adaptation to and mitigation of, the negative consequences of climate change.