IARC/JAXA Terrestrial Team Workshop - February 22, 2006

The role of fire disturbance in the response of historical carbon dynamics in the boreal forest from 1950-2002

M. S. Balshi & A. D. McGuire

In the boreal forest, wildfire is a common occurrence, and changes in the fire regime have consequences for carbon dynamics as well as water and energy feedbacks to the climate system.  Changes in climate and atmospheric CO₂ concentrations may also affect carbon dynamics through their effects on ecosystem processes.  However our ability to project future temporal and spatial changes in carbon dynamics is limited by our understanding of how the temporal and spatial aspects of fire influence historical carbon dynamics. 
To evaluate the temporal and spatial changes of carbon dynamics in response to CO₂, climate, and fire disturbance, we developed a fire module for the Terrestrial Ecosystem Model (TEM) and simulated carbon dynamics for the pan-boreal region north of 45° N from 1950-2002.  We conducted three simulations: CO₂ fertilization only, CO₂ and climate variability, and CO₂, climate, and fire disturbance.  For fire simulations, information on historical fire return interval (FRI) was used for backcasting fire disturbance prior to the start of the historical fire records.  We used cokriging estimates based on data for the IGBP high latitude transects in Eurasia and estimated FRIs for North America based on spatially and temporally explicit fire records for the period 1950-2002. Simulation results for the pan-boreal region north of 45° N indicate that C storage increased in response to CO₂, climate, and fire at a rate of 344 Tg C yr^-1 between 1950 and 2002.  Partitioning the effects of CO₂, climate, and fire for North America indicates that from 1950-2002, atmospheric CO₂ was responsible for sequestering 37.52 Tg C yr^-1 (3.48 g C m^-2 yr^-1 ), climatic variation was responsible for sequestering 38.09 Tg C yr^-1 (3.54 g C m^-2 yr^-1 ), and fire was responsible for releasing 7.01 Tg C yr^-1 (0.62 g C m^-2 yr^-1 ).  For Eurasia, atmospheric CO₂ was responsible for sequestering 126.31 Tg C yr^-1 (4.96 g C m^-2 yr^-1 ), climatic variation was responsible for sequestering 70.87 Tg C yr^-1 (2.78 g C m^-2 yr^-1 ), and fire was responsible for sequestering 78.64 Tg C yr^-1 (3.08 g C m^-2 yr^-1 ).  Our analysis suggests that CO₂, climate, and fire each play important roles in carbon dynamics across the pan-boreal region.  It also shows that it is important to incorporate fire in a temporally and spatially explicit manner when estimating the effects of fire on carbon dynamics for the boreal forest region.  Our next step in this study is to develop a fire model that can be coupled to TEM to evaluate carbon dynamics across the boreal forest for future scenarios of climate change.

Fire effects on ecosystem services

Terry Chapin

Climate warming has had large, societally important effects on Alaska. Ecosystem services, which are the benefits that society derives from ecosystems, provide one way to evaluate these effects. I briefly discuss the effects of climate warming on ecosystem services in Alaska, with an emphasis on the effects associated with changes in wildfire. Warming has had particularly profound effects on factors that influence landscape interactions (climate regulation, disturbance spread, and disease regulation). Ecosystem goods, such as food (subsistence resources), water, and wood that receive most management attention are only indirectly affected by warming. The cultural services provided by ecosystems are also sensitive to warming and have led to some of the few institutional responses that address causes of climate warming.

Measurement of root respiration and soil respiration before and after forest fire – for evaluation of the role of root respiration in soil respiration

Masako Dannoura, and Mayuko Jomura

CO₂ efflux from belowground is an important role in carbon cycling of boreal ecosystems due to the high proportion of biomass allocated belowground. Soil respiration consists of root respiration (autotrophic respiration) and decomposition respiration (heterotrophic respiration), and to evaluate root respiration separately is necessary to understand the carbon cycling. We established two plots in heavy burnt area and control area at Pocker Flat. Soil respiration was measured at both plot using automatic chamber system by IRGA (same as Dr. Jomura, but the bottom of the chamber which is suitable to measure soil respiration). Root respiration was measured by root sampling, and the sample root size was classified to 0^-2 , 2-5, 5-20, 20-50, 50< mm in diameter. Measurements were conducted in 8th to 13th Aug. 2005. Air temperature was 5 to 35 degrees centigrade in this term.
At heavy burnt area, water content of root was low and root respiration was very low (almost zero). At control plot, the smaller root had the higher respiration rate per weight. The root respiration per sample surface area was 0.05-0.08 (mgCO₂ m^-2 s^-1 ). The value was similar to root respiration measured in temperate forest in Japan under same temperature. Soil respiration at heavy burnt area was 0.01-0.04 (mgCO₂ m^-2 s^-1 ) and it was 30-60% of that of control plot. But soil respiration at control plot was lower than that of temperate forest. So, the ratio of root respiration to soil respiration was higher than temperate forest at least in summer season. In heavy burnt area, root respiration was hardly included in soil respiration. It is suspected that soil respiration will be higher with production of root. We will estimate root respiration per area and evaluate the role of the root respiration including soil respiration.

download full paper

Ground Surface Burn Severity Index based on Thermal Conductivity Measurements

Masami Fukuda
Institute of Low Temperature Science, Hokkaido University

It is important to evaluate the burned severity related to the impacts of thermal regime of permafrost and ecosystems.  At Poker Flat, the profiles of thermal conductivity in upper layers were obtained by means of in situ measurements.  The residual organic layer depth is determined by the profiles of thermal conductivity measurements.  The difference of thermal conductivity between unburnt and burnt sites is about one order magnitude. Therefore using thermal conductivity profile at each site, the severity of burn is clearly defined.  At the most severely burnt site in Poker Flat, only a few centimeter of organic layer was detected.  The following impact of loss of organic layer is deepening of active layer.  Some estimation of active layer change is conducted.

download full paper

Are decreases in snow cover moderated by increased carbon storage in fire-disturbed high-latitude terrestrial ecosystems?

E.S. Euskirchen¹, A.D. McGuire², M.S. Balshi¹, F.S. Chapin¹

1. Institute of Arctic Biology, University of Alaska Fairbanks
2. U.S. Geological Survey, Alaska Cooperative Fish and Wildlife Research Unit

High-latitude terrestrial ecosystems play an important role in the earth’s climate system due to the broad expanse that is occupied by fire-disturbed vegetation and seasonally snow-covered ground.  As snow retreats in response to increasing temperatures in these regions, less solar energy is reflected into space and more energy is absorbed and transferred to the atmosphere.  This results in a positive snow/albedo feedback loop that reinforces warming.  This warming may be moderated by the enhanced capacity of these terrestrial ecosystems to sequester carbon under changes in atmospheric CO₂ concentrations, climate, and fire regimes.  We compared these responses retrospectively based on simulations with a large-scale terrestrial ecosystem model for the land area north of 50º N.  Our analysis took into account two historical 30-year time periods, 1920-1940 and 1970-2000, where surface air temperatures generally increased and snow cover generally decreased.  Decreases in snow cover duration from 1920-1940 were approximately 0.9 –1.7 days decade^-1 , and were primarily due later snowfall in the autumn.  From 1970-2000, the trend in snow cover duration was greater, decreasing by 1.6^-3 .8 days decade^-1 , generally caused by earlier snow melt in the spring.  Across the entire study domain, our findings suggest that changes in energy due to changes in snow cover show a heating effect of +1.1 W m^-2 during 1920-1940 with this trend increasing to +1.9 W m^-2 between 1970-2000.  In comparison, changes in energy due to changes in atmospheric CO₂ concentrations, climate, and fire regimes showed a cooling effect of -0.17 W m^-2 between 1920 – 1940 and ^-1 .0 W m^-2 between 1970-2000.  These results indicate that the effects of a longer snow-free season on atmospheric energy balances should considered in studies of climate change, particularly with respect to associated shifts in vegetation between forests, grasslands, and tundra.

Preliminary report on 2005 and 2004 Alaska forest fires

Hiroshi Hayasaka

In 2004 and 2005, many large-scale forest fires had occurred in Alaska. The burnt area in 2004 was 26,142 km² and it was 18,816 km² in 2005. They were the first and the third historical record since 1956. The cause of these fires is now under investigating. This is a preliminary report on these fires. In 2004, severe lightning in June and July had caused most fires. Not a few fires grew into large-scale fires with the help of severe drought and Chinook or foehn phenomena. As a result, the total burnt area reached about 26,000 km² and itis almost the same area of the Lake Erie. One of the large-scale fires called “Boundary fire” occurred near Fairbanks was chosen to investigate fire growth process in detail. Boundary fire has been considered as a second largest fire in 2004. Fire growth of large-scale forest fire was clearly showed by some analytical results of the hot spot (fire), climate, and fire history data. “Boundary fire” and other fires became a large-scale fire for the following processes.

1. Sever lightning occurred in the beginning of June and ignited various places in the boreal forest in Alaska. One of these lightning ignited the forest in “Boundary fire” area on 13th June and first hot spot was detected on 18th June.
2. First hot spot peak appeared on 30th June due to the dry weather conditions that made by Chinook.
3. Most fires were self-extinguished or lost activeness due to the large massive smoke from severe fires.
4. Second hot spot peak was found on 13th July and made by drought from the beginning of June.
5. Most fires in the second peak were extinguished due to the rainfall of the end of July.
6. Third hot spot peak appeared on 11th August due to the drought from the beginning of August.

download full paper

Survey results of ground conditions after fires near Fairbanks

Hiroshi Hayasaka and Hirosi Tani

In 2004 and 2005, many large-scale forest fires had occurred in Alaska. Many severe damages occurred in Alaska, nevertheless this is a good chance for scientists to investigate various phenomena arose as a result of forest fire from various points of view. Before large fires occurred, only a few burnt areas can be accessible by car from Fairbanks. But after large fires, we can easily visit several burnt sites. 2004 Boundary fire, 2005 Fish Creek fire and 200? Eagle Summit were chosen as investigation areas. In these forest fire or burnt areas, thickness and weight measurement in burnt or unburned forest floor, sampling of vegetation, situation observation and counting of stand and fallen trees in severely burnt area, surface temperature measurement on burnt and unburned forest floors, smoke observation near actual fire site, observation of slow flame propagation at actual fire site, observation of forest fire site in steep slop, observation of landslide due to forest fire in steep slop and observation of highland tundra fire were carried out. One spot of Poker flat area within 2004 Boundary fire region were investigating intensively by more than ten Japanese researchers of various research fields from 2005. We also joined this intensive field investigation. Two 20 x 20 m research areas were chosen in unburned and partially burnt respectively. Each research areas were also divided into four areas.  Measurement and sampling were carried out in the center of each four divided areas. Thickness and weight measurement were done by making 20x20 cm size hole in the forest floor at the center of each divided areas. Sample of vegetation, mainly moss litter and duff, were chosen from 20x20 cm cut blocks. As a preliminary result of measurement, thickness of moss litter in partially burnt area is 40% thinner than that of unburned area. Carbon content (t/ha) of moss litter in partially burnt area is 23% smaller than that of unburned area.

Impacts of Wildfire on the Hydrological Environment in Interior Alaska

Yoshiyuki Ishii¹, Yuji Kodama¹, Yong-Won Kim², Koichiro Harada³, Yuki Sawada¹, and Masami Fukuda¹

1. Institute of Low Temperature Science, Hokkaido University, Japan
2. International Arctic Research Center, University of Alaska Fairbanks, USA
3. Department of Environmental Sciences, Miyagi University, Japan

“Boundary Fire 2004” in the boreal forest of Interior Alaska was the largest wildfire in these 50 years, and it is predicted to affect strongly to the hydrology, permafrost degradation, and vegetation recovery in the watershed. We made hydrological observations to examine the impacts of the wildfire in May and August of 2005 at the north-facing slope in the Poker Flat Research Range (GI/UAF), 50 km northeast of Fairbanks, Alaska. In a heavily-burned headwater basin, rainfall-runoff response of the small stream was reduced from May to June. These changes could be caused by the increase of soil water storage volume with the increase of thawing layer thickness. At the moss burned site on the hillslope, frost table was deepening to 1- 1.5 m during the summer. However, it was kept nearly the same, 0.4 m in deep, at the control site. This indicates the high insulation effect of the moss, and its burnout increases the active layer thickness significantly.

download full paper

Decomposition rate of woody debris in a burnt forest: Results of a preliminary study at Poker flat research range.

Mayuko Jomura, Masako Dannoura

Wildfires consume large amount of carbon not only by the combustion of tree and moss biomass but also the decomposition of burnt materials and soil organic matter after fire. Increase in the fire in boreal forest caused by climate change will stimulate microbial decomposition of organic matter and forest soil and consequentially boreal forest may become a net source of carbon. Thus, investigation of the dynamics of postfire decomposition is particularly important. In this study, we focused on postfire microbial decomposition of organic matter and examined control factors and characteristics of decomposition processes.  This study examined the decomposition respiration of woody debris (WD) in a black spruce forest at Poker flat research range. The forest was burned in June 2004. We measured decomposition respiration of WD (RWD) using closed dynamic chamber system with infrared gas analyzer and temperature and water content of WD in August 2005. WD samples (diameter: 3-10cm) were obtained from standing dead wood (snag) and downed dead wood (log) of black spruce.  Temperature of WD was high (about 25°C), nevertheless RWD was very low (n=10, snags: 0.21, logs: 0.40 mgCO₂ kg^-1 h^-1 ). If this environmental condition continues year-round, decomposition rate of snags and logs was 0.001 and 0.002 y^-1 , respectively and mean residence time of snags and logs was about 1000 and 500 years. The low decomposition rate may be mostly induced by the extremely low water content of WD (both snags and logs: 0.18g g^-1 ). The slope facing to the south, well-drained soil, the lack of the crown of living trees and little precipitation may cause soil drying in the experimental site in summer. Similarly, WD water content became low resulting in low microbial activity. In spring, snow melt and temperature increase may stimulate microbial activity and decomposition. Thus, to determine decomposition dynamics of WD after forest fire, seasonal changes in decomposition rate should be measured. Moreover, because of only one year after the forest fire, microbes did not entered sufficiently in WD and this may be the initial stage of decomposition.  RWD was different between snags and logs, however, water content was almost similar. This difference was induced by low RWD of snag samples located at the high position (more than 4m) and some of these samples did not expose CO₂. Thus, the height of the WD position affects microbial invasion resulting in low decomposition rate. The vertical position of WD may affect decomposition rate of WD due to both the difference in microbial invasion and water content of WD. Therefore, the vertical position of WD may be a significant factor to determine decomposition dynamics of WD.

download full paper

Assessing the Severity of the 2004 Alaskan Fires through Satellite and Field Observations

Eric S. Kasischke and Elizabeth Hoy, Department of Geography, University of Maryland, College Park, Maryland
Merritt R. Turetsky, Departments of Plant Biology/Fisheries & Wildlife, Michigan State University, East Lansing, Michigan
A. David McGuire, Department of Biology and Wildlife, University of Alaska,  Fairbanks, Alaska
Nancy H.F. French, Altarum Institute, Ann Arbor, Michigan

We are just beginning the second year of two NASA funded research projects whose focus is on developing approaches to estimate the amounts of carbon released during the burning of surface organic layers in black spruce forests and peatlands that are common throughout the North American boreal forest. The objectives of these studies are fourfold: (a) to quantify the variability in surface fuel consumption that occurs during boreal fires in North America; (b) to understand the factors that cause this variability; (c) to assess using information derived from satellite imagery to map variations in surface characteristics that can be related to surface and aboveground fire severity; and (d) to incorporate our improve understanding of factors resulting in variations in burn severity into the Terrestrial Ecosystem Model to examine how variations in the North American fire regime have influenced terrestrial carbon source-relationships in this region (see paper by Balshi et al.).

During the first year of our study, we focused on: (a) conducting field studies in the fires that took place in interior Alaska in 2004 and 2005 to collect data to measure the variability in surface fuel consumption in black spruce forests and to assess fire severity using different indices (such as the composite burn index); and (b) assessing the relationship between the composite burn index and the normalized burned ratio derived from Landsat imagery. Results from these studies will be presented, along with our plans for further field studies during 2006 will be presented.

Observation of soil CO₂ efflux at Poker Flat forest fire burned site in 2005

Yuji Kodama (ILTS, HU), Yoshiyuki Ishii (ILTS, HU), Yonwon Kim (IARC)

Forest fire gives tremendous changes to surface conditions.  In order to clarify the difference in soil respiration and its response to weather and soil conditions between burned and unburned, measurement using chambers were carried out in the summer of 2005.  Together with the respiration measurement, chamber and soil temperatures and soil moisture as well as the meteorological conditions were observed. Soil CO₂ efflux showed a clear diurnal variation at the both burned and unburned site.  It was correlated well with soil temperature and the difference between the measured and the predicted by soil temperature was correlated with soil moisture.  Albedo of the burned site was very small (4-7%) and it was increased with the amount of grown vegetation (10-17%).  Soil temperature was low where soil moisture was large.  

download full paper

Fire effects on plant-soil system in taiga forests in interior Alaska

Lina Koyama¹, Muneto Hirobe², Satoru Hobara³ and Naoko Tokuchi⁴

1. Graduate School of Informatics, Kyoto University, Kyoto, Japan
2. Graduate School of Environmental Science, Okayama University, Okayama, Japan
3. Rakuno Gakuen University, Hokkaido, Japan
4. Field Science Education and Research Center, Kyoto University, Kyoto, Japan

Nutrient availability is assumed to increase and stimulate ecosystem productivity by global warming in many terrestrial ecosystems. Northern forests contain significant global carbon pools, and fire is a common component in these forests. Thus, concerns have recently been directed to fire effects on carbon and nitrogen cycles in plant-soil system. Our objective of this study is to clarify fire effects on 1) surface soil properties, especially soil N transformations, 2) dissolved organic matter characteristics in soil, 3) plant N use and 4) openness of N cycle. In Aug. 2005, a preliminary research was carried out in the Poker Flat Research Range located in the heavily burned area by a large forest fire in 2004. Pool size of inorganic N, metals, and Dissolved Organic Carbon, and net rates of N mineralization and nitrification were compared (1) between burned and unburned areas in three different vegetation stands (black spruce, black spruce/paper birch, and aspen) and (2) between inside and outside of unburned “moss island”. Soil inorganic N pool size was larger in the burned stands than in the unburned stands, while rate of net N mineralization (measured using laboratory incubation) was greater in the unburned stands than in the burned stands. Between inside and outside of “moss island”, no clear difference was found in both inorganic N pool and net N mineralization rate, despite lower temperature in “moss island”. Fire effects on the change of vegetation are partly due to the species physiological characteristics about nutrient use. For example, results of a fertilization experiment in a tundra ecosystem showed that the responses of plant to N fertilization were highly variable among species, and some species increased their biomass by N fertilization but not N concentration, and the other species tended to be the opposite. These results and next research plans are discussed using our results on plant N use and openness of N cycle studied in Central Siberia and Alaskan tundra.

download full paper

Spectral characteristics of ground components one year after fire in an interior Alaskan black spruce forest

Keiji Kushida
Institute of Low Temperature Science, Hokkaido University, Japan

Our final goals are to understand component spectral characteristics of black spruce fire chronosequence in interior Alaska, and to model and evaluate detectability of vegetation changes after fire and according carbon budget changes from remotely sensed data. In 2005, the spectral reflectances in 350 – 2500 nm were measured at 80 points (155 samples) in an interior Alaskan black spruce forest (Poker Flat site), which was burned in the summer of 2004, and analyzed the spectral separativeness of the representative ground components (burnt sphagnum mosses, damaged sphagnum mosses, live sphagnum mosses). The points were situated in or around the sixteen 10 m × 10 m plots established by Tsuyuzaki et al. For observing damaged sphagnum mosses and live sphagnum mosses, surface undergrowths on the mosses of the measurement points were removed. The spectral reflectances at all of the points were measured under entirely diffuse illumination conditions. When the solar illumination was specular, an artificial shadow was made on the objects and the reference panel. The spectral reflectances of 21 cases were observed under both specular and diffuse illuminations. As a result, we obtained spectral characteristics of burnt sphagnum mosses, damaged sphagnum mosses, and live sphagnum mosses, and the three had a significant difference. Further, the arial ratio of the ground componsnts were estimated by using the spectral characteristics and Landsat ETM+ imagery (resolution: 15 m – 30 m) taken on 4 Aug. 2004. The results can be used for base information to interpret MODIS (250 m – 1 km), and ALOS (2.5 m – 10 m) satellite data.

download full paper

The role of tire on the surface energy balance at a sub-arctic tussock tundra site

Anna Liljedahl, Larry Hinzman, Kenji Yoshikawa, Robert Busey
Water and Environmental Research Center
International Arctic Research Center

Niagara Creek (6.5km²), situated at central Seward Peninsula, Northwestern Alaska, was affected by a severe burn August 2002. Soil and meteorological observations were made before and after the fire at a fixed location, with vegetation consisting of Eriphorum tussocks, year 2000 to 2005. During this time period, the annual averaged soil temperature increased 2.5±0.6ºC throughout the 1 m deep profile as a positively correlated increase with time following the burn. Local noon energy flux ratios during the second and third summer following fire indicate an altered energy partitioning in comparison to prefire situations, whereas the radiation efficiency (net radiation normalized to incoming short wave radiation) stays relatively stable. Summer 2004 exhibits noon albedo values of0.13±0.01 compared to prefire data of 0.17±0.01 and 0.16±0.01. Unstable atmospheric conditions are found more prone to occur after the burn, represented by the Richardson number. Near surface soil display enhanced post-fire soil moisture levels following spring melt, close to saturation, throughout the thawed season, a phenomenon still present summer 2005. Thermokarst formation and severe erosion occurred along Niagara Creek streambed after the fire, which resulted in a formation of a > 3500 m³ void, an estimation made fall 2005. Hydrological postfire model simulations of Niagara Creek watershed do not exhibit a large difference from the calibrated hydrograph, with only slightly higher peak flows. Evapotranspiration is reduced during post-fire simulations, while no clear change in recession periods after rainstorm events was found.

Carbon loss from forest floor/top soil by wild fire: a case study of 2004 fires at Poker Flat Research Range.

Y. Matsuura

Forestry and Forest Products Research Institute, Tshukuba, Japan

Carbon loss after one year of 2004 wild fire was estimated in Poker Flat Research Range (PFRR). By comparing organic C storage at burned site with those of unburned sites, C loss from forest floor/top soil was estimated at three different forest types. Three profiles were surveyed at Picea mariana (black spruce) stands, of which two profiles were located on the burned forest. Heavily burned site showed remarkable permafrost table subsidence below 160cm of mineral soil horizon; on the other hand, permafrost table existed at the depth of 33 cm below mineral soil at the unburned black spruce stand. Forest floor thickness of unburned black spruce stand was 20 to 30 cm. Decline of forest floor thicknesses after wild fire in Betula papyrifera var. humilis (Alaskan paper birch) stands was distinct with slightly scorched top soil, whereas there was not so much decline of forest floor thickness in Populus tremuloides (quaking aspen) stands. Organic C loss from forest floor was estimated as follows; 2.0 – 2.2 kg C m^-2 in black spruce stands, 2.6 kg C m^-2 in Alaskan paper birch stands, 0.2 kg C m^-2 in quaking aspen stands. C loss form top soil (30 cm storage) was not so clear in birch and aspen stands. Further studies on dead root organs and CWD contribution to C storage and loss are needed.

download full paper

Forest fire detection based on MODIS satellite imagery, and Comparison of NOAA satellite imagery with fire fighters' information.

Koji Nakau, Masami Fukuda, Keiji Kushida, Hiroshi Hayasaka, Keiji Kimura, Hiroshi Tani

Impacts of boreal forest fires have absorbed intensive attention because of huge fires in these years in Alaska as well as Siberia.  To reduce impacts of forest fire in boreal forest area, the early fire detection is one of essential components in firefighting activity because of difficulties of fire suppressing in remote area without water.  Here, we developed fire detection information system from receiving AVHRR to output fire detection map and validated the early detection algorithm using AVHRR satellite imagery.  Forest fires were detected using an algorithm; two-dimensional histogram method by Prof. Kudo.  This algorithm uses a threshold on mid-infrared band 3 and a two-dimensional histogram of visible band 1 and thermal infrared band 5 as a looking up table; these detection criteria corresponds becoming to burnt to black, thermal emission by burning.  As a ground truth data, we collected reports of fires observed by local firefighters in Siberia and reports of JAL passenger flights.  We compared satellite detected pixels with location of reported fires.  We aggregated this comparison by fires to estimate the fire detection rate and early fire detection rate.  We found the fire detection rate was surprisingly different between fires reported by firefighters and by passenger flights.  Finally, we found the reason of the different fire detection rate as scale of fires observed.  This implies difficulty on forest fire detection especially for small sized forest fires, and also implies the importance of ground truth data especially reported by fire fighters.  We are planning extend the area collecting ground truth data delivered from local firefighting agencies in Alaska and Siberia from this summer season to validate the forest fire detection algorithms using AVHRR and MODIS.  As a preparation, we made a system to detect forest fires every day automatically for area of entire Alaska last summer.

download full paper

Effects of fire on fine root biomass in a black spruce forest: A preliminary study at the Poker Flat Research Range.

Kyotaro Noguchi and Yojiro Matsuura
Forestry and Forest Products Research Institute, Tsukuba, Japan

Fine root is a key component in nutrient cycling of forested ecosystems.  In this study, fine root mass (< 2 mm in diameter) were examined at a severely burned black spruce forest at the Poker Flat Research Range in August 2005, one year after severe wildfire in summer 2004.  An unburned black spruce forest close to the burned site was also investigated as a control.  Estimated fine root mass, including live and dead fine roots, in surface moss + organic layer to the soil depth of 20 cm was ~1010 and ~1040 g m^-2 in burned and unburned sites, respectively, 70% of which was concentrated to surface moss + organic layers in both sites.  However, because of technical difficulties, we could not separate roots from root-organic matter complexes in part and they were 360 and 900 g m^-2 in burned and unburned sites, respectively.  Thickness of the surface moss + organic layer was smaller in the burned site (~22 cm) than in the unburned site (~38 cm), which resulted in larger fine root density at the surface layer in burned site (~4.4 kg m^-3 ) than in unburned site (~2.4 kg m^-3 ).  Roughly estimated living proportion of fine roots was 1% and 65% in the burned and unburned plots, respectively.  Although there are technical problems to be solved, the results of this study suggested that wildfire in 2004 substantially affected ground surface condition (decline in moss + organic layer) and killed most of fine roots at the severely burned black spruce forest, but the effects on total fine root mass might not be remarkable. 

download full paper

Examination of structural constraints in relation to site condition in black spruce chronosequence

Osawa, A.¹, T. Kajimoto², Y. Matsuura³ & N. Kurachi⁴

1. Faculty of Intercultural Communication, Ryukoku University, Ohtsu, Japan
2. Kyushu Research Center, Forestry and Forest Products Research Institute, Kumamoto, Japan,
3. Forestry and Forest Products Research Institute, Tsukuba, Japan,
4. Hiraoka Forest Institute, Ohtsu, Japan

A hypothesis that a forest ecosystem growing over permafrost has a relatively small value of the maximum attainable aboveground biomass which is specific to the site condition of the forest will be tested for black spruce (Picea mariana) forests developing after stand-replacing forest fires in Alaska and adjacent Yukon and Northwest Territories, Canada.  Also tested will be the truncation of the relationship between aboveground biomass and stand density characterized as the self-thinning rule at a relatively young stage of stand development in black spruce.  We recently noted that these hypotheses may apply to Gmelin larch (Larix gmelinii) forests of central Siberia, and intend to examine if the same is observed among the black spruce ecosystems in North America.  Site condition of any forest will be estimated as the mean tree height of the stand for a given stand age, applying the Russian system of site index.  Depth of the soil active layer will be measured in late summer or early fall, and is compared to the estimated site index and stand biomass.  Partitioning of fixed organic matter among various organs, growth, and death of individual trees, and patterns of leaf biomass and three-dimensional leaf distribution in the stand and their changes over time will be described with a use of chronosequence to infer mechanisms of the structural stand development in black spruce. 

download full paper

Modeling Climate-Fire-Vegetation Dynamic

S. Rupp

The boreal forest version of ALFRESCO was developed to explore the interactions and feedbacks between fire, climate, and vegetation in interior Alaska.  ALFRESCO is a state-and-transition model of successional dynamics that explicitly represents the spatial processes of fire and vegetation recruitment across the landscape.  ALFRESCO does not model fire behavior, but rather models the empirical relationship between growing-season climate (e.g., average temperature and total precipitation) and total annual area burned (i.e., the footprint of fire on the landscape).  ALFRESCO also models the changes in vegetation flammability that occurs during succession through a flammability coefficient that changes with vegetation type and stand age. We will report on several research projects utilizing ALFRESCO to simulate current and future landscape dynamics including response to forecast climatic warming and the influence of fire frequency on caribou habitat.

Assessment of Remotely Sensed Index for Mapping Burn Severity in Interior Alaska’s Black Spruce Forests

Andrew Ruth and David Verbyla, University of Alaska Fairbanks

Alaska experienced two large fire seasons in 2004 and 2005 with over 6.7 and 4.5 million acres burned, respectively. A large portion of these burns occurred in black spruce (Picea mariana) ecosystems of Interior Alaska. Land managers and fire management officials are especially interested in the effects of burn severity on future stand trajectories in black spruce ecosystems because of issues related to wildlife habitat improvement and natural fuel breaks. However, the remoteness and scale of Alaskan fires prevent feasible ground or aerial truthing of burn severity. Burn severity assessments in Alaska’s black spruce communities are potentially amenable to remote sensing. Based on research in Alaska and the lower-48, there is a strong linear relationship between a remotely sensed index, the Differenced Normalized Burn Ratio (dNBR) and a field-based metric, the Composite Burn Index (CBI).  However, the strength of this relationship has not been specifically tested within the black spruce communities of Interior Alaska and our goal is to assess the correlation within this one vegetation type.

Our objectives are:
Determine the association between a remotely sensed burn severity index (dNBR) and a field-based burn severity metric (CBI) within black spruce communities.
Assess the sources of variation in the remotely sensed burn severity index including:
    a) Date of pre and post-fire imagery
    b) Canopy versus soil burn severity field estimates

Permafrost Degradation After the Tundra Fire in Seward Peninsula, Alaska - A perennial Study-

Yuki Sawada, Institute of Low Temperature Science, Hokkaido University, Japan
Koichiro Harada, Miyagi University, Japan

Thermal, water and electrical conditions of permafrost after the tundra fire were observed in Seward Peninsula, southwest Alaska, in order to evaluate the effect of fire on permafrost conditions. Field observations were made in 2005 and four sites were established where the slope direction and surface disturbance condition are different; south- or north-facing, and burned or unburned. At each site ground temperature and water content were measured by pit survey, and the seasonal thawed depth measurements were also conducted by using the steel rod from the ground surface. Transient electromagnetic surveys were carried out along profiles with the length of 140-180m to compare the permafrost condition using a transmitter loop of 60 x 60m. The temperatures of 20-40cm deep at the burned sites were 4-5 ºC higher than that at the unburned sites. The soil water contents at the burned sites showed the high condition. The measured thawed depths are significantly different between the burned and unburned sites, which were more than 20cm deeper in the burned sites than that in the unburned sites. The obtained apparent resistivity curves and estimated resistivity models showed that a significant difference was observed between south- and north-facing slopes. At the north-facing sites, high resistivity layers were estimated near the ground surface with the thickness of 20-26m, which represents permafrost. The permafrost base could not be detected at the south-facing sites because the base is located in bedrock. There is no significant difference of the curves and models between burned and unburned sites. However, only at the burned south-facing site, stable data could be obtained by using the standard central induction configuration, which means that this site has a relative low resistivity condition near the ground surface. Thus, the burned south-facing site may have a different permafrost condition near the surface.

download full paper

Preliminary study on permafrost degradation by 2004 Boundary fire: Site description, active layer, and geomorphological control on the vegetation recovery

Koichiro Harada, Miyagi University, Japan
Yuki Sawada, Institute of Low Temperature Science, Hokkaido University, Japan
Junko Mori, Graduate School of engineering, Hokkaido University

2004 Alaskan wildfire occurred near Fairbanks (The “Boundary fire”) heavily burned surface organic matters and soils. Surficial soil loss may cause permafrost degradation, and it finally affects slope stability, water circulation regime, and vegetation recovery. Under the inter-discipline observation program, we attempt to clarify 1) Changes on permafrost distribution, 2) Changes on thermal regime of permafrost / active layer, 3) Topographical changes due to the soil loss and permafrost degradation. In the first summer, measurements of thermal regime in the melting layer and microtopography were carried out. Pit excavation in August revealed that the removal of organic layer was a major controlling factor for the melting depth. Melting depth was deepest (>1m) in the heavily-disturbed sites (H), while the shallowest depth (<0.5m) appeared in the less-disturbed sites (L). In the moderately-disturbed sites (M), melting depth drastically varied in short distance, due to the mosaic-like distribution of the remained organics and Sphagnum. Under the thick organic layers in M sites, melting depth was very shallow (<0.5m), appearing similar to the melting depth in the L sites. Photogrammetric measurements on 1m quadrats highlighted very rough micro-topography on the heavy burned site. Digital images were corrected by a consumer-level digital camera, and were analyzed with photogrammetry software and GIS. The final contour map and DEM have a resolution of less than 1mm. Small ridges and troughs exhibited on the burned surface with relative height of approx. 5 – 10cm. These small ridge and troughs seems to determine the distribution of seedling. 2D coordinates of seedlings on the quadrats, which were measured by Tsuyuzaki and Narita, agree well with the position of troughs. This suggests that the subsequent changes of micro-topography after the fire disturbance may control the distribution of seedlings.

download full paper

Initial watershed response to boreal forest fires in Interior Alaska

Horacio Toniolo

The summer of 2004 in Alaska was characterized by enormous and devastating boreal forest fires. Small streams draining water from areas affected by fires in different proportions (i.e., unburned, partially, and severally burned) were systematically sampled during the summer of 2005. All the streams were located in watersheds underlain by discontinuous permafrost. In order to collect daily water samples, autosamplers were deployed in the streams after spring breakup. Pressure transducers and dataloggers in conjunction with velocity measurements were used to estimate water discharge in the streams. Human influence is negligible in the study areas, with the exception of modifications caused by fire suppression activities. Thus, collected data from these areas can be considered as a natural system response to forest fires.  Initial data will be presented in the meeting. 

go back

The effects of intensive forest fire on revegetation in interior Alaska (mid-term report, February 2006)

TSUYUZAKI Shiro, Graduate School of Environmental Earth Science, Hokkaido University
NARITA Kenji, Faculty of Education and Human Studies, Akita University

To detect the effects of large-scaled fire on the revegetation of Picea mariana forest,we set up 16 10 m × 10 m plots at Poker Flat near Fairbanks, Alaska, USA, in the spring of 2005. Forest fire occurred in this region in the summer of 2004. Owing to the fire, stemdensity declined 9%-100% and canopy openness increased in the plots surveyed. The groundcover mostly consisting of Sphagnum was burned by the fire, and remained patchily. Burnedground surface in the plots ranged from 3% to 100%. The relationship between the height ofsurvived trees and age determined by tree core samples was positively and linearly correlated.The frequency of tree stems gradually decreased with increasing tree height. Those resultssuggested that tree recruitment had gradually occurred so far. We set up six 1 m × 1 mquadrats in each plot, and recorded plant cover on each species and marked all seedlings ineach quadrat. Of vascular plants, small shrubs, such as Betula nana and Ledumgroenlandicum, and sedges (Carex spp.), that recovered vegetatively, were common on theunburned ground surface, while Epilobium angustifolium were common on burned surface.In addition, even on the burned ground surface, shrubs, e.g., Betula nana and Ledumgroenlandicum, and perennial sedges, such as Carex bigelowii, could survive vegetativelythroughout the fire with low cover. We found out the safe sites for seedling emergence variedgreatly between tree species. Picea mariana germinated on Sphangum mat while Betulapapyrifera and Populus tremuloides emerged on bare ground where the aboveground coverincluding duff was completely removed by the fire. The surveys on the relationships betweenrevegetation patterns and its related environmental factors will be continued.

download full paper

go back

Heterotrophic respiration declines following wildfire in black spruce soils of interior Alaska.

David Valentine, Evan Kane, and Tim Quintal
University of Alaska, Fairbanks AK, United States

Fire creates conditions that stimulate microbial activity in and carbon loss rates from many forest soils.  We compared post-fire soil respiration rates in two pairs of burned and unburned black spruce forests in contrasting topographic positions.  The upland sites experienced a moderate burn in July 1999 during the Frostfire experimental wildfire, and the lowland were burned more severely during the “Survey Line” wildfire in May 2001.  Both had unburned sites nearby that were suitable for comparison.  Heterotrophic respiration was measured using root exclusion collars in unburned soils.  Using darkened chambers two years following their respective fires, we found growing season soil respiration rates were higher in the unburned lowland black spruce site (peak of 275 mg C/m2/h) than in the unburned upland site (peak of 200 mg C/m²/h).  In both, peak season burned soil respiration rates were about 60% lower than unburned.  Heterotrophic respiration in burned soils was significantly slower than in unburned soils.  This decline was small in the upland soils, but averaged 32% in the lowland stands.  These results contradict the widely held assumption of a stimulating effect of fire.  We attribute the negative impact of fire on the loss of root and other detrital inputs to soil until vegetation re-establishes.

Impacts of wildfire on the permafrost in the boreal forests and the tussock tundra

The impact to the permafrost during and after wildfire was studied using multi year fire sites including two-controlled burns. Heat transfer by conduction to the permafrost was not significant during fire. Immediately following fire ground thermal conductivity may increase 10-fold depending upon the extent of burning of the surfical organic soil. The thickness of the remaining organic layer strongly affects permafrost degradation and aggradation. If the organic layer thickness was not reduced during the burn, then the active layer did not change after the burn, in spite of the surface albedo decrease.  Any significant disturbance to the surface organic layer will increase heat flow through the active layer into the permafrost.  Approximately three to five years after severe disturbance and depending upon site conditions, the active layer will increase to a thickness that does not completely refreeze the following winter.  This results in formation of a talik.  Model studies suggest that if an organic layer of more than 7-12 cm remains following a wildfire, then the thermal impact to the permafrost will be minimal in the boreal forests of Interior Alaska. Tussock tundra contains possibly different heat transfer system (non-conductive heat transfer) that may cause more cooling of the ground after the fire.