- Email: [email protected]
Boreal lichen woodlands: A possible negative feedback to climate change in eastern North America
Agricultural and Forest Meteorology 151 (2011) 521–528
Contents lists available at ScienceDirect
Agricultural and Forest Meteorology journal homepage: www.elsevier.com/locate/agrformet
Boreal lichen woodlands: A possible negative feedback to climate change in eastern North America P.Y. Bernier a,∗ , R.L. Desjardins b , Y. Karimi-Zindashty b , D. Worth b , A. Beaudoin a , Y. Luo c , S. Wang d a
Canadian Forest Service, Natural Resources Canada, Québec, QC, Canada Agriculture and Agrifood Canada, Ottawa, ON, Canada Meteorological Service of Canada, Environment Canada, Ottawa, ON, Canada d Canada Centre for Remote Sensing, Natural Resources Canada, Ottawa, ON, Canada b c
a r t i c l e
i n f o
Article history: Received 24 September 2010 Received in revised form 15 November 2010 Accepted 25 December 2010 Keywords: Albedo Carbon sequestration Reforestation Fire regime
a b s t r a c t Because of successive forest fires, closed-canopy black spruce forests are susceptible to a shift towards open lichen–spruce woodlands in parts of the boreal forest of eastern North America. The shift from dark black spruce canopies to pale lichen ground cover offers a dramatic contrast in reflectance that may compensate for the CO2 emissions from forest fires in terms of radiative forcing. We have therefore looked at the climate change feedback that would result from the generation of lichen woodlands through changes in albedo and in stored carbon. Using albedo estimates based on MODIS imagery and incoming solar radiation for the period between 2000 and 2008 along with forest biomass estimates for eastern Canada, we have estimated that net radiative forcing for the conversion from closed-canopy coniferous forests to open lichen woodlands would be about −0.12 nW m−2 ha−1 , and would therefore generate a cooling effect in the atmosphere. Based on current estimates of area in open lichen woodlands within the closed-canopy black spruce–moss forests of eastern Canada, we estimate that a current net forcing of −0.094 mW m−2 has already arisen from such conversions. As projections of future climate have been linked to increased probability of forest fires, the generation of open lichen woodlands provides a possible negative feedback to climate change. Results also suggest that carbon sequestration through the afforestation of boreal lichen woodlands may not provide a climate change mitigation benefit. Crown Copyright © 2011 Published by Elsevier B.V. All rights reserved.
1. Introduction The boreal forest is Canada’s largest forest ecosystem, occupying 35% of the total Canadian land area and 77% of Canada’s total forest land (Natural Resources Canada, 2010). Fire regularly disturbs the boreal forest in Canada, leaving behind a temporary landscape of burnt trees and blackened earth. On average, about 9000 forest fires consume 2.1 million hectares annually in this ecosystem. While destructive, forest fires are a necessary and natural aspect of the life cycle of the boreal forest, releasing nutrients and rejuvenating the forest. The dominant tree species of the boreal forest are adapted and can regenerate following fires. Aspen (Populus tremuloides) regenerates vegetatively and abundantly from its root system following a fire, while black spruce (Picea mariana) and jack pine (Pinus banksiana Lamb.) have serotinous or semi-serotinous cones and heat-tolerant seeds that ensure a rapid re-seeding of the
∗ Corresponding author at: Canadian Forest Service, Natural Resources Canada, PO Box 10380, Stn. Ste-Foy, Québec, QC, Canada G1V 4C7. E-mail address: [email protected] (P.Y. Bernier).
ground where these species are present (Elliott-Fisk, 2000). Black spruce is a slow-growing species that can take 25 years to reach sexual maturity (Morneau and Payette, 1989). In eastern Canada, where pure stands of this species dominate the high-latitude boreal forest, a first burn or harvest followed by a second burn before trees reach sexual maturity may therefore eliminate or strongly reduce the tree cover, depending on the fire intensity. In welldrained landscapes of coarse glacial deposits, the now open ground often becomes covered with various species of lichens, mostly of the Cladina and Cladonia genus, creating stable open lichen woodland ecosystems that persist over time (Girard et al., 2008). Each land cover type has its own surface reflectivity, or albedo. When snow free, the boreal forest has an albedo of 9–11%, which tends to be less than that of other forest types and grasslands, which usually is in the 10–20% range. This difference in albedo is accentuated in winter, as snow blankets landscapes with sparse forest cover, resulting in an albedo in the range of up to 78%, but will tend to fall through a densely forested landscape, leaving the absorptive branches exposed and resulting in an albedo in the range of 10–26% (Betts, 2000; Betts et al., 2007a). Therefore, a change in land cover type has a direct effect on albedo, which in turn influences the local
0168-1923/$ – see front matter. Crown Copyright © 2011 Published by Elsevier B.V. All rights reserved. doi:10.1016/j.agrformet.2010.12.013
522
P.Y. Bernier et al. / Agricultural and Forest Meteorology 151 (2011) 521–528
radiation budget. This change can be expressed in terms of radiative forcing, which is defined as the disturbance of net irradiance at the tropopause after allowing the stratospheric temperature to readjust to radiative equilibrium (Chapter 6 in IPCC, 2001). Radiative forcing is often used to combine the effects of different agents, such as greenhouse gases, with a change in albedo into one metric that describes the net impact on climate. For instance, Betts (2000) showed that the radiative forcing of carbon sequestration and albedo change associated with afforestation/reforestation could be compared, and Pielke et al. (2002) estimated that for the boreal forest of North America this forcing ranged from 0.05 to 0.2 nW m−2 ha−1 . At the global level, a nominal climate sensitivity of 0.75 ◦ C per W m−2 of radiative forcing on a global scale has been reported by Hansen et al. (2008). Air temperature is increasing within the boreal region of Canada (Jones and Moberg, 2003), which may lead to an increase in the future area of forest burnt annually (Flannigan et al., 2005). In the boreal forest of eastern Canada, the burn rate since 1940 has decreased compared with historical rates (Le Goff et al., 2007), but predictions for a 2 × CO2 climate scenario suggest a 38% increase in future burn rates compared with current rates across the coniferous forest of Québec (Bergeron et al., 2006). It has been speculated that a climate change-induced increase in forest fires could lead to a positive climate feedback as carbon dioxide (CO2 ), methane (CH4 ) and nitrous oxide (N2 O) released during combustion promote further warming and burning (Kurz et al., 1995; Metsaranta et al., 2010). However, the larger impact may be through the change in surface albedo following a burn, particularly in early spring when solar radiation is increasing and snow still blankets recently burnt areas, whereas in forested areas most of the snowfall passes through or falls off the canopy, leaving the more absorptive branches exposed (Betts and Ball, 1997). Such a difference in albedo may result in a net cooling during winter (Bonan et al., 1992; Brovkin et al., 2004; Claussen et al., 2001; Ganopolski et al., 2001). Over the multiple decades required for forest regrowth, simulations suggest that the negative feedback of albedo could outweigh the positive feedback effects of CO2 emissions from the combustion of organic matter (Randerson et al., 2006). In the boreal forest of eastern Canada, the change in albedo following repeated fires may be particularly important because the ground cover of open lichen woodlands is very pale in colour with high albedo, and because these open systems are stable and not easily recolonised by trees. The increased rate of creation of open lichen woodlands through projected increases in fire frequency may therefore provide a negative feedback to climate change. The objectives of the present work were to evaluate the net radiative forcing resulting from the loss of biomass carbon to the atmosphere and the increase in albedo resulting from the creation of open lichen woodlands in the closed-canopy forest domain, and to assess the potential net radiative forcing effect of increased lichen woodland extent as a result of increased fire frequency. Results are also interpreted in terms of afforestation of open lichen woodlands.
2. Materials and methods 2.1. Study location The study area is located in the boreal forest of the province of Québec, Canada, and consists of the closed-canopy black spruce–moss bioclimatic domain (Robitaille and Saucier, 1998), located roughly between 49◦ N and 50◦ 30 N latitude, and covering an area of about 412,400 km2 (Saucier et al., 2009). This area is characterised by a low mean annual temperature (−2.5 ◦ C to 0.0 ◦ C), with annual precipitation increasing from west to east, from about 700–1200 mm. The dominant tree species is P. mariana (black
Fig. 1. Location of the reference blocks selected for the local dominance of one of four target cover densities of black spruce–lichen stand types. The darker shaded area corresponds to the black spruce–moss bioclimatic domain of the Quebec forest ecological classification of Robitaille and Saucier (1998).
spruce) with a variable abundance of P. banksiana (jack pine) and P. tremuloides (quaking aspen) in the dryer western portion of the domain, and of Abies balsamea (balsam fir) and Betula papyrifera (paper birch) in the wetter eastern portion of the domain. With the notable exception of the James Bay lowlands, most of these landscapes are underlain by coarse glacial deposits draped over hilly terrain. The historical fire return interval for this region varies across the landscape from about 140 years in the west to 280 years or more in the east. The tendency since 1940 has been a decrease in the fire frequency, with a doubling of the fire return interval to nearly 600 years in the east (Bergeron et al., 2006; Le Goff et al., 2007). Simulations done for 2 × CO2 and 3 × CO2 scenarios suggest a return to historical burn rates and possibly an increase in frequency above the historical rates for the western portions of the domain, with return intervals of 80 years for the 3 × CO2 scenario (Flannigan et al., 2005). Lichen woodlands are scattered in portions of this bioclimatic domain and have been expanding into the closed-canopy boreal forest for the past 1500 years. A possible cause of this expansion could be the drier late-Holocene conditions for this region (Jasinski and Payette, 2005), although fire risk seems to have been declining in the eastern boreal forest over the past 7000 years (Hély et al., 2010). 2.2. Selection of reference blocks Albedo values were determined for four forest types with a lichen ground cover, and with different canopy cover densities: 0–10%, 10–25%, 25–40% and 40–60%. The computation of albedo started with the selection of reference blocks for each cover density on a forest cover map that was developed as part of the Earth Observation for Sustainable Development (EOSD) project led by the Canadian Forest Service and funded by the Canadian Space Agency (Wulder et al., 2008). Location of the reference blocks is shown in Fig. 1. The circa 2000 land cover classification of Canada was completed in 2007 based on Landsat TM and ETM multispectral scenes. Within this national framework, the Enhancement-Classification Method (ECM) (Beaubien et al., 1999) was applied over the province of Québec using three Landsat spectral bands (red, NIR and MIR) from an assembly of 100 Landsat scenes acquired between 1998 and 2003. ECM was used to produce a detailed land cover map of Québec at a 25-m resolution with 52 thematic classes. These classes were further aggregated into the 23 EOSD thematic classes to be compliant with the national-level EOSD legend (Wulder et al., 2008). In our study, we used the detailed version of the land cover map of Québec to depict the forest covers of interest, i.e.
P.Y. Bernier et al. / Agricultural and Forest Meteorology 151 (2011) 521–528
523
Table 1 Number and area of EOSD-based reference blocks of lichen woodlands (EOSD ID 19 and 31) and of black spruce–lichen forests (EOSD ID 29 and 28) used for delineating MODIS pixels for albedo estimations, as well as the relative importance of these forest types within the 412,400 km2 black spruce–moss bioclimatic domain. Also shown is the number of field plots that were obtained for each forest cover density class from the EOSD plot database, and the mean values (with standard deviation in parenthesis) of total tree biomass in these plots. EOSD ID
19 31 29 28
Cover density class (%)
0–10 10–25 25–40 40–60
Number of reference blocks
7 5 4 4
Total area sampled (km2 )
17 32 3 77
conifer-dominated (75%+ of conifer species, mostly black spruce) with a lichen-dominated ground cover, and representing forest cover density classes of 0–10%, 10–25%, 25–40% and 40–60% (Table 1). Based on a regional accuracy assessment exercise conducted in the James Bay area in 2006 of the national-level EOSD land cover product and using a similar protocol as Wulder et al. (2006), overall classification accuracy was 73%, whereas specific class accuracy was 92%, 66% and 72%, respectively, for the 0–10%, the combined 20–25% and 25–40%, and the 40–60% density classes (without taking into account the different types of moss cover). More than 97% of the misclassified pixels were found in the immediate-neighbour density classes. Based on these accuracy figures, we can reasonably state that the accuracy for the four classes in this study is at least better than 75% considering the distinctive spectral signatures of black spruce stands with highly reflective lichen ground cover. For each cover density class of the black spruce–lichen forest type, we identified and extracted from four to seven small representative landscape blocks distributed across the black spruce–moss domain of Québec (Table 1). Each block had to have a minimum size of 0.5 km2 in order to include at least eight 250-m MODIS pixels, and had to be dominated by the forest cover type of interest. These blocks provided the location of nearly pure stands of the forest covers of interest for multi-spectral determination of albedo from the 250-m resolution MODIS images. The registration error for the 250-m MODIS pixel is usually less than 50 m (Wolfe et al., 2002), assuring a good spatial match between selected EOSD forest cover blocks and the MODIS pixels. Forest canopies present a continuum of cover density. However, experience in the photo-interpretation of forest cover shows that below a certain threshold tree cover ceases to be continuous and adopts a clumped structure. In commercial forestry, this threshold is set at 25% of the cover density class, where 25% is the lower bound of the last cover density class deemed to be potentially harvestable. Cover densities below 25% are simply classified as open woodlands. In the current study, we chose to adopt the same threshold between closed-canopy forests and open lichen woodlands and consider the two cover density classes of 0–10% and 10–25% to be lichen woodlands, while the two other denser classes are considered to be closed-canopy forests. Another element of comparison is the lichen-dominated ground cover. The most representative forest type in the black spruce–moss bioclimatic domain is that of pure black spruce stands with ground cover mosses such as Pleurozium schreberi (BSG) and Hylocomnium splendens (Hedw.). Mosses are dark green and will confer a lower albedo to the stand than white or grey lichen ground species such as Cladonia spp. and Cladina spp. For the present exercise, we chose to look only at the forest types with a lichen ground cover. By doing so, we consider only the change in albedo induced by the change in forest cover density and can therefore have a more direct comparison with radiative forcing associated with the carbon sequestered in trees.
Portion of bioclimatic domain (%)
1.9 3.1 4.6 1.6
Biomass
Number of plots
Mean (sd) (t ha−1 )
11 54 41 24
18.0 (9.8) 42.4 (19.5) 61.4 (21.8) 102.4 (30.5)
2.3. Solar radiation data Downward shortwave radiation was measured as a 30-min average at the Québec Flux Station–Eastern Old Black Spruce (49.693◦ N, 74.342◦ W), a CO2 flux measurement site of the Fluxnet Canada/Canadian Carbon Program (Margolis et al., 2006). This station has been operating continuously since 2003. Solar radiation is measured using a Kipp and Zonen CNR1 net radiometer, which measures the four component fluxes of net radiation. In the shortwave spectrum, the CNR1 integrates over the spectral range of 305–2800 nm. The CNR1 was mounted on a horizontal boom at a height of 22 m, approximately 8 m above a moderate density black spruce-dominated canopy with an understory of mosses and ericaceous shrubs (see Bergeron et al., 2007 for additional description of the site). The 30-min solar radiation values were used to produce monthly values of daily average incoming solar radiation over the 2003–2008 data record. As MODIS data were available for albedo determination for the longer 2000–2008 period, we used the 2003–2008 mean monthly values of daily solar radiation to match the 2000–2002 albedo data. Similar radiation measurements were also made at a companion site where trees had been harvested in 2000. Both radiation records were used to compute albedo, referred to below as “field albedo”, to be used for verification of MODIS albedo estimates. 2.4. Albedo data Surface albedo values were estimated using the Moderate Resolution Imaging Spectroradiometer (MODIS) located on board the TERRA satellite, available since 2000. The MODIS instrument provides global coverage every 1–2 days in 36 spectral bands ranging from visible to infrared and to thermal wavelengths between 405 and 14,385 nm. For the current exercise, in terms of albedo, we used the first seven spectral bands (B1–B7) with wavelength coverage of 459–2155 nm, which are specifically designed for land shortwave radiative applications. Two of them, B1 (620–670 nm) and B2 (841–876 nm), acquire imagery at a 250 m spatial resolution at nadir, while the rest (B3–B7) have a 500-m resolution. Bands B3–B7 were therefore downscaled to a 250-m resolution using a regression and normalization scheme for compatibility with B1 and B2. We first used bands B1 and B2 as inputs to generate the surrogate intermediate 250-m resolution images for each band from B3 to B7. The values in these images were then normalized to the observed values from the 500-m imagery in order to preserve the radiometric properties of the original data. Clouds and cloud shadows were detected and classified at a 250-m resolution by investigating multi-spectral features. The 10-day clear-sky composite data were then projected onto a standard LCC (Lambert Conformal Conic) map (Luo et al., 2008). A 10-day long interval is a practical compromise for MODIS to likely obtain clear-sky observations and to quickly capture any change in surface properties. The clear-sky composites of B1–B7 reflectance were then used to produce spatially
524
P.Y. Bernier et al. / Agricultural and Forest Meteorology 151 (2011) 521–528
continuous spectral albedo as in Luo et al. (2005). In that method, albedo is computed using a BRDF (Bi-directional Reflectance Distribution Function) model adjusted to observed reflectance from various angles but from the same land cover type. The modified Ross-Thick Li-Sparse BRDF model was chosen for its good performance in MODIS albedo retrieval (Luo et al., 2005). The seven spectral albedo data sets were then converted into one broadband albedo data set using the MODIS polynomial approximation conversion of Liang et al. (1999). This albedo value represents the overall solar short-wave albedo. The reference blocks selected from the EOSD land cover map for the four target forest types were remapped onto the LCC projection map. Their 25-m pixels were then aggregated into 250-m resolution pixels that matched the MODIS 250-m resolution pixels. Each aggregated pixel was assigned to the forest type that was dominant within the 25-m pixels used in the aggregation. For each block, pixels of the four forest types were grouped according to canopy cover densities and the albedo values from each group were averaged. The exact number of pixels for a given group within a given reference block varied over time because of cloud contamination or failure in albedo retrieval. Averaging was done only when there were at least four pixels available. For each month we obtained three albedo values from the three 10-day composite intervals, and therefore 27 values per month for the 2000–2008 period of record from which we computed the monthly albedo means and standard deviations. Multi-year monthly mean (and standard deviations) values of albedo per forest cover type were computed from the 27 monthly values obtained for the 2000–2008 record.
2.5. Forest carbon data
shortwave radiation in W m−2 (a) was estimated as: a = I(1 − ˛) + ˇ(˛I)
where I is measured downward shortwave radiation (W m−2 ), ˛ is albedo, and ˇ is a coefficient to correct for the atmospheric absorption of reflected shortwave radiation (Montenegro et al., 2009; Ramanathan et al., 1987; Weaver et al., 2001). The value of ˇ originally set to 0.3 by Weaver et al. (2001) was lowered to 0.23 by Montenegro et al. (2009) to account for atmospheric absorption of both downward and upward shortwave radiation. The net reflected shortwave radiation escaping the atmosphere is the result of the difference between incoming and absorbed radiation, which can be calculated as follows: NRS = I − a
(2) (W m−2 ).
where NRS is the net reflected shortwave radiation The differences in reflected shortwave radiation for various land cover types affect the local radiation budget. This effect on global radiative flux can be estimated for a given region of interest by: ıR =
(NRS1 − NRS2 )Ar Ae
2.6. Radiative forcing We calculated the radiative forcing attributable to change in albedo for each 250 m resolution pixel for the different forest densities as in Montenegro et al. (2009) and Betts (2000). Absorbed
(3)
where Ar is the area of the region of interest, Ae is the Earth’s surface area, and subscripts 1 and 2 refer to the different land cover types. Using this equation, the radiative forcing attributable to different land cover types can be determined. Biomass carbon contents obtained from the field plots of different forest cover densities were used to compute RFC , the equivalent radiative forcing, in nW m−2 ha−1 if this carbon was released into the atmosphere (Betts, 2000; Montenegro et al., 2009):
RFC = 5.35 ln 1 + Forest inventories had been carried out across the northern portion of the boreal forest for the explicit purpose of relating forest biomass to multi-spectral imagery and EOSD land cover classes in order to map the biomass of non-inventoried boreal forests (Beaudoin et al., 2007). Areas sampled (and average longitude and latitude of the plot locations) were Chibougamau (49.69◦ N, 74.34◦ W), Mistassini (50.90◦ N, 74.73◦ W), Radisson (55.55◦ N, 76.40◦ W), and Wabush (52.92◦ N, 66.99◦ W). Additional plots were measured in the James Bay lowlands of the black spruce–moss bioclimatic domain, but the mostly clay deposits and flat landscapes of these areas produce stands that follow different developmental dynamics than those from other portions of the domain. These plots were therefore left out. A total of 144 plots were retained for the analysis, with observed cover density values ranging from 5% to 85%. The tally of tree diameter at breast height and species in each plot was used to compute total aboveground biomass using the species-specific allometric equations of Lambert et al. (2005). The coarse root fraction was estimated at 18% of total biomass (Czapowskyj et al., 1985). Carbon concentration in the biomass was assumed to be 50%. Carbon in the non-biomass pools was left out in the present analysis as it was assumed that non-biomass carbon would be constant across the range of forest cover densities in the black spruce–lichen forest types. This assumption is supported by soil carbon data from open-canopy and closed-canopy black spruce and jack pine stands with lichen ground cover (Lagacé-Banville, 2009).
(1)
C C0
(4)
where C0 is the atmospheric CO2 concentration, and C is the change in atmospheric CO2 concentration if the carbon in the forest biomass is added to the atmosphere. Present day atmospheric CO2 concentration was assumed to be 385 ppmv (Menon et al., 2010). A value of RFC was thus obtained for the four cover density classes of black spruce–lichen forest types considered in the present study. 3. Results and discussion 3.1. Radiation and albedo data Downward solar radiation data show a strong seasonality typical of high-latitude locations (Fig. 2), with multi-year monthly average maximum and minimum daily downward shortwave radiation of 223 and 36 W m−2 recorded in June and December, respectively. Monthly average albedo values determined from MODIS observations for different forest densities (Fig. 3) show the expected increase in albedo with decreasing forest cover density. Albedo values are also highest when snow is on the ground, from December to April, with a maximum in February, and are lowest in summer. Albedo values start decreasing in late winter before the appearance of any bare ground resulting from the dirtying of the surface following initial melting events (Langham, 1981) and the faster discharge of snow from branches. Snow also enhances the difference in albedo among the different forest cover densities (Fig. 3). Snow-free monthly average albedo values of 0.16, 0.13, 0.11 and 0.09 for areas with less than 10%, 10–25%, 25–40% and 40–60% forest cover densities, respectively, are comparable to other published values. Betts and Ball (1997) report average summer albedo values of 0.2, 0.15 and 0.083 for areas covered by grass, aspen and conifer, respectively. Betts (2000) reports values of 0.18–0.21 for arable crop land and of 0.14–0.15 for dense evergreen coniferous forest under snow-free conditions. Furthermore, for deep snow
P.Y. Bernier et al. / Agricultural and Forest Meteorology 151 (2011) 521–528
525
Fig. 2. Multi-year monthly average of downward shortwave radiation for the 2003–2008 period at Fluxnet Canada/Canadian Carbon Program’s Québec Flux Station–Eastern Boreal Black Spruce. The error bar represents one standard deviation.
conditions, he reports a maximum albedo value of 0.26 for densely forested areas, which is somewhat lower than our winter maximum of 0.31 for the 40–60% canopy cover forest type. Monthly mean field albedo measurements obtained at the forested and harvested sites of the Québec Flux station bracketed the albedo estimates from the four target forest types, except in late spring to late summer when the albedo value of the harvested site was lower or equal to that of the lichen woodland with the lowest cover density (0–10%) (results not shown). 3.2. Radiative forcing Radiative forcing as a result of differences in albedo was computed as the difference in reflected shortwave radiation (Eq. (1)) between our highest forest cover density (40–60%) and each of the three other forest densities. Negative values reflect a reduction in absorbed solar radiation, and hence a cooling effect. Mean annual radiative forcing values of −0.26, −0.12 and −0.06 nW m−2 ha−1
were estimated for the forest cover types of 0–10%, 10–25% and 25–40% in comparison with 40–60% coverage, respectively (Fig. 4). The greatest radiative forcing for all cover densities occurred during the snow-covered season of January–April and peaked in mid-April (Fig. 4) because of the combination of pre-melting highalbedo snow (Fig. 5) and increasing shortwave radiation (Fig. 2). As expected, the most negative value of shortwave radiative forcing was obtained for the 0–10% cover density. This estimated albedo effect is not quite as pronounced as the 0.35 nW m−2 ha−1 calculated by Betts et al. (2007b) for albedo change resulting from afforestation, likely on account of differences in forest cover properties between the studies. Radiative forcing as a result of CO2 emissions following fire-driven loss of biomass was computed as the difference in RFC , the warming potential of the mean standing biomass if released into the atmosphere (Eq. (4)), between the 40–60% forest cover density type and the three lower cover density types. A positive value indicates a warming potential for the atmosphere. Mean annual values of radiative forcing of 0.13, 0.10
Fig. 3. Multi-year monthly averages of albedo determined for four forest cover types of different cover densities, with lichen ground cover, for the 2000–2008 period. The error bars represent one standard deviation.
526
P.Y. Bernier et al. / Agricultural and Forest Meteorology 151 (2011) 521–528
Fig. 4. Multi-year monthly average of radiative forcing effect of conversion from closed-canopy coniferous forest (40–60% cover density) to lower density closed-canopy forest and to open lichen woodlands for the 2000–2008 period. The error bar represents one standard deviation.
and 0.080 nW m−2 ha−1 were estimated for the forest cover types of 0–10%, 10–25% and 25–40% in comparison to the 40–60% forest cover type, respectively (Fig. 6). The largest value represents the impact of a change from closed-canopy forest to open lichen woodland, as observed following successive burns in that boreal environment. The net radiative forcing linked to the fire-driven creation of lower density forest canopies in these spruce–lichen environments is then computed as the sum of the albedo- and biomass-derived radiative forcings. Results from this computation (Fig. 6) suggest a net cooling effect of −0.12 nW m−2 ha−1 when the cover density drops from the 40–60% class to the 0–10% canopy density cover type, and an approximately neutral effect for a reduction from the 40–60% to the 10–25% and 25–40% cover types. These results suggest that the creation of very open lichen woodlands through repeated burning of closed-canopy forests has a net cooling effect on the atmosphere. The corollary to these results is that reforestation of open lichen woodlands and the creation of closed-canopy
forests may have at best no effect on global radiative forcing, or at worse a negative impact of 0.12 nW m−2 ha−1 for the conversion of 0–10% to 40–60% cover density forests. Currently, according to the EOSD-based classification (Table 1), open lichen woodlands in the 0–10% cover density class cover 1.9% of the black spruce–moss bioclimatic domain in the province of Québec, which represents an area of 7835 km2 . Therefore, gradual conversion of closed-canopy black spruce forests to lichen woodlands has to date generated a radiative forcing of −0.094 mW m−2 . However, it is at the moment difficult to estimate the current rate of lichen woodland generation throughout this domain, let alone predict how climate change will affect it. Given the functional relationship between successive disturbances and the generation of open lichen woodlands, these results nevertheless suggest a potential for an increased negative feedback between climate change and the boreal forest. Other studies have also suggested a warming effect due to afforestation because of changes in albedo (e.g. Rotenberg and
Fig. 5. Multi-year monthly averages of snow depth (m) for the 2003–2008 period at Fluxnet Canada/Canadian Carbon Program’s Québec Flux Station–Eastern Boreal Black Spruce.
P.Y. Bernier et al. / Agricultural and Forest Meteorology 151 (2011) 521–528
527
Fig. 6. Net radiative forcing resulting from the conversion from closed-canopy coniferous forest (40–60% cover density) to lower density closed canopy forest and to open lichen woodlands for the 2000–2008 period. The error bar represents one standard deviation.
Yakir, 2010). Of particular interest for our study, Betts (2000) calculated an increasing net warming effect due to afforestation with increasing latitudes in the boreal forests of eastern Canada. However, the very coarse resolution results in mixtures of land covers in each pixel, including the high proportion of natural and man-made water bodies in the boreal forest of Canada, thus making it difficult to attribute effects to specific forest cover types. Montenegro et al. (2009), on the other hand, suggest that high latitude afforestation/reforestation always results in a net cooling effect. Their study, however, was confined to agricultural areas, with high latitude calculations encompassing mostly areas in Eurasia with milder and drier climates and less snow cover, as well as ground cover types that may have albedo values lower than that of lichen. Their results are therefore valid for afforestation of land currently under cultivation, but may not be applicable to our present example which compares closed-canopy black spruce forests and open lichen woodlands. Our study specifically targets forest types along a gradient of forest cover densities down to open lichen woodlands, and is limited to land areas. It should therefore offer a more accurate evaluation of changes in albedo following a specific fire-induced forest cover change in the boreal environment.
Although albedo and carbon release and sequestration are likely first-order effects of land cover changes on the climate system, the total effect comprises more complex second-order interactions. Basic hydrological research has shown that the removal of trees reduces evapotranspiration (Bosch and Hewlett, 1982) and thus latent heat transfer, which can lead regionally to a reduction in cloud cover and of top-of-atmosphere albedo (Betts et al., 2007a). Forests also emit more longwave radiation than cooler high-albedo surfaces, but this radiation is reabsorbed by the atmosphere to varying degrees depending on the exact wavelength of the radiation. Ideally, all such effects should be incorporated into a common metric to compare the impact of land cover changes on the climate system, but this is still a very challenging task (see Lee, 2010; Leu, 2010; Yakir and Rotenberg, 2010). The issue is partially one of scale, with case studies looking at ground-level effects of local land cover changes whereas total radiative forcing must consider the larger effects on the global climate system. In areas where land cover drives socio-economic activities, impacts of land cover changes on energy use by society must also be included in such calculations. This is clearly an area in need of additional work. Acknowledgements
4. Conclusion Based on the computation of albedo changes and the impact of carbon release or sequestration on the atmospheric greenhouse effect, we conclude that the conversion of closed-canopy boreal forests to lichen woodlands through repeated fires has a cooling effect on the atmosphere of about −0.12 nW m−2 ha−1 , and that conversion to date has generated a total radiative forcing of −0.094 mW m−2 . Since climate change may enhance the natural conversion rate through increased fire frequency combined with forest harvesting, the additional generation of lichen woodlands would produce a negative feedback to climate change. Conversely, the conversion of open lichen woodlands into closedcanopy forests may generate a warming effect to the atmosphere. Calculations done with other types of boreal open woodlands, such as those dominated by ericaceous shrubs, may reveal a positive effect of reforestation. These results are important in their own right, but they also emphasize the need to incorporate changes in albedo in the evaluation of activities related to land use, land use change and forestry for climate change mitigation.
The authors thank the Fluxnet Canada/Canadian Carbon program for providing high-quality radiation data, as well as Stephen Côté and Philippe Villemaire of the Canadian Forest Service, Natural Resources Canada, for extracting reference blocks from the EOSD land cover map. Thanks are also extended to Dr. Alan Betts and to one anonymous reviewer for their constructive comments during the review process. The MODIS data were acquired from the NASA Goddard Space Flight Center, Level 1, and the Atmosphere Archive and Distribution System (LAADS) (http://ladsweb.nascom.nasa.gov). Part of the funding for this study was provided by the Agriculture and Agri-Food Canada program on Sustainable AGricultural Environmental Systems (SAGES). References Beaubien, J., Cihlar, J., Simard, G., Latifovic, R., 1999. Land cover from multiple thematic mapper scenes using a new enhancement-classification methodology. J. Geophys. Res. 104 (D22), 27900–27920. Beaudoin, A., Guindon, L., Hall, R.J., Luther, J.E., Fournier, R.A., 2007. A strategy to map biomass of Canadian northern boreal forests using multi-sensor/multi-
528
P.Y. Bernier et al. / Agricultural and Forest Meteorology 151 (2011) 521–528
resolution imagery: approach, methods and preliminary results. In: Proceedings of FORESTSAT 2007 Forests and Remote Sensing: Methods and Operational Tools , November 5–7, 2007, Montpellier, France. Bergeron, O., Margolis, H.A., Black, T.A., Coursolle, C., Dunn, A.L., Barr, A.G., Wofsy, S.C., 2007. Comparison of carbon dioxide fluxes over three boreal black spruce forests in Canada. Global Change Biol. 13 (1), 89–107. Bergeron, Y., Cyr, D., Drever, C.R., Flannigan, M., Gauthier, S., Kneeshaw, D., Lauzon, È., Leduc, A., Le Goff, H., Lesieur, D., Logan, K., 2006. Past, current, and future fire frequencies in Quebec’s commercial forests: implications for the cumulative effects of harvesting and fire on age-class structure and natural disturbancebased management. Can. J. Forest Res. 36, 2737–2744, doi:10.1139/X06-177. Betts, R.A., 2000. Offset of the potential carbon sink from boreal forestation by decreases in surface albedo. Nature 408, 187–190. Betts, A.K., Ball, J.H., 1997. Albedo over the boreal forest. J. Geophys. Res. 102 (D24), 28901–28909. Betts, A.K., Desjardins, R.L., Worth, D., 2007a. Impacts of agriculture, forest and cloud feedback on the surface energy budget in BOREAS. Agric. Forest Meteorol. 142, 156–169. Betts, R.A., Falloon, P.D., Klein Goldewijk, K., Ramankutty, N., 2007b. Biogeophysical effects of land use on climate: model simulations of radiative forcing and largescale temperature change. Agric. Forest Meteorol. 142, 216–233. Bonan, G.B., Pollard, D., Thompson, S.L., 1992. Effects of boreal forest vegetation on global climate. Nature 359, 716–718. Bosch, J.M., Hewlett, J.D., 1982. A review of catchment experiments to determine the effect of vegetation changes on water yield and evapotranspiration. J. Hydrol. 55, 3–23. Brovkin, V., Sitch, S., Von Bloh, W., Claussen, M., Bauer, E., Cramer, W., 2004. Role of land cover changes for atmospheric CO2 increase and climate change during the last 150 years. Global Change Biol. 10, 1253–1266. Claussen, M., Brovkin, V., Ganopolski, A., 2001. Biogeophysical versus biogeochemical feedbacks of large-scale land cover change. Geophys. Res. Lett. 28, 1011–1014. Czapowskyj, M.M., Robison, D.J., Briggs, R.D., White, E.H., 1985. Component biomass equations for black spruce in Maine. USDA Forest Serv. Res. Paper NE-564. Elliott-Fisk, D.L., 2000. The taiga and boreal forest. In: Barbour, M.G., Dwight Billings, W. (Eds.), North American Terrestrial Vegetation. , second ed. Cambridge University Press, Cambridge, pp. 41–74. Flannigan, M.D., Logan, K.A., Amiro, B.D., Skinner, W.R., Stocks, B.J., 2005. Future area burned in Canada. Climatic Change 72, 1–16. Ganopolski, A., Petoukhov, V., Rahmstorf, S., Brovkin, V., Claussen, M., Eliseev, A., Kubatzki, C., 2001. CLIMBER-2: a climate system model of intermediate complexity. Part II: Model sensitivity. Clim. Dynam. 17, 735–751. Girard, F., Payette, S., Gagnon, R., 2008. Rapid expansion of lichen woodlands within the closed-crown boreal forest zone over the last 50 years caused by stand disturbances in eastern Canada. J. Biogeogr. 35, 529–537. Hansen, J., Sato, M., Kharecha, P., Beerling, D., Berner, R., Masson-Delmonte, V., Pagani, M., Raymo, M., Royer, D.L., Zachos, J.C., 2008. Target atmospheric CO2 : where should humanity aim? Open Atmos. Sci. J. 2, 217–231. Hély, C., Girardin, M.P., Ali, A.A., Carcaillet, C., Brewer, S., Bergeron, Y., 2010. Eastern boreal North American wildfire risk of the past 7000 years: a model-data comparison. Geophys. Res. Lett. 37, L14709, doi:10.1029/2010GL043706. IPCC (Intergovernmental Panel on Climate Change), 2001. Climate Change 2001: The Scientific Basis. Contribution of Working Group I to the Third Assessment Report of the Intergovernmental Panel on Climate Change. Cambridge University Press, Cambridge. Jasinski, J.P., Payette, S., 2005. The creation of alternative stable states in the southern boreal forest, Québec, Canada. Ecol. Monogr. 75, 561–583. Jones, P.D., Moberg, A., 2003. Hemispheric and large-scale surface air temperature variations: an extensive revision and an update to 2001. J. Climate 16, 206–223. Kurz, W.A., Apps, M.J., Stocks, B.J., Volney, W.J.A., 1995. Global climate change: disturbance regimes and biospheric feedbacks of temperate and boreal forests. In: Woodwell, G.M., Mackenzie, F.T. (Eds.), Biotic Feedbacks in the Global Climatic System: Will the Warming Feed the Warming? Oxford University Press, Oxford, pp. 119–133. Lagacé-Banville, J., 2009. Caractérisation des stocks de carbone de 5 types de formations végétales dans un secteur du bassin versant de la rivière Eastmain, Baie James. M.Sc. Thesis, Université du Québec à Montréal, 159 pp. Lambert, M.-C., Ung, C.-H., Raulier, F., 2005. Canadian national tree aboveground biomass equations. Can. J. Forest Res. 35, 1996–2018.
Langham, E.J., 1981. Physics and properties of snow cover. In: Gray, D.M., Male, D.H. (Eds.), Handbook of Snow: Principles, Processes Management and Use. Pergamon Press, Oxford. Le Goff, H., Flannigan, M.D., Bergeron, Y., Girardin, M.P., 2007. Historical fire regime shifts related to climate teleconnections in the Waswanipi area, central Quebec, Canada. Int. J. Wildland Fire 16 (5), 607–618. Lee, X., 2010. Forests and climate: a warming paradox. Science 328, 1479. Leu, S., 2010. Forests and climate: the search for specifics. Science 328, 1479–1480. Liang, S., Strahler, A.H., Walthall, C., 1999. Retrieval of land surface albedo from satellite observations: a simulation study. J. Appl. Meteorol. 38, 712–725. Luo, Y., Trishchenko, A.P., Khlopenkov, K.V., 2008. Developing clear-sky, cloud and cloud shadow mask for producing clear-sky composites at 250-meter spatial resolution for the seven MODIS land bands over Canada and North America. Remote Sens. Environ. 112, 4167–4185. Luo, Y., Trishchenko, A.P., Latifovic, R., Li, Z., 2005. Surface bidirectional reflectance and albedo properties derived using a land cover-based approach with Moderate Resolution Imaging Spectroradiometer observations. J. Geophys. Res. 110, D01106, doi:10.1029/2004JD004741. Margolis, H.A., Flanagan, L.B., Amiro, B.D., 2006. The Fluxnet-Canada Research Network: influence of climate and disturbance on carbon cycling in forests and peatlands. Agric. Forest Meteorol. 140, 1–5. Menon, S., Akbari, H., Mahanama, S., Sednev, I., Levinson, R., 2010. Radiative forcing and temperature response to changes in urban albedos and associated CO2 offsets. Environ. Res. Lett. 5, 014005, doi:10.1088/1748-9326/5/1/014005. Metsaranta, J.M., Kurz, W.A., Neilson, E.T., Stinson, G., 2010. Implications of future disturbance regimes on the carbon balance of Canada’s managed forest (2010–2100). Tellus B, doi:10.1111/j.1600-0889.2010.00487.x. Montenegro, A., Eby, M., Mu, Q., Mulligan, M., Weaver, A.J., Wiebe, E.C., Zhao, M., 2009. The net carbon drawdown of small scale afforestation from satellite observations. Global Planet. Change 69, 195–204. Morneau, C., Payette, S., 1989. Postfire lichen-spruce woodland recovery at the limit of the boreal forest in northern Quebec. Can. J. Bot. 67, 2770–2782. Natural Resources Canada, 2010. The Atlas of Canada. http://atlas.nrcan.gc.ca/site/english/index.html. Pielke Sr., R.A., Marland, G., Betts, R.A., Chase, T.N., Eastman, J.L., Niles, J.O., Niyogi, D.D.S., Running, S.W., 2002. The influence of land-use change and landscape dynamics on the climate system: relevance to climate-change policy beyond the radiative effect of greenhouse gases. Philos. Trans. R. Soc. A 360, 1705– 1719. Ramanathan, V., Callis, L., Cess, R., Hansen, J., Isaksen, I., Kuhn, W., Lacis, A., Luther, F., Mahlman, J., Reck, R., Schlesinger, M., 1987. Climate–chemical interactions and effects of changing atmospheric trace gases. Rev. Geophys. 25 (7), 1441–1482. Randerson, J.T., Liu, H., Flanner, M.G., Chambers, S.D., Jin, Y., Hess, P.G., Pfister, G., Mack, M.C., Treseder, K.K., Welp, L.R., Chapin, F.S., Harden, J.W., Goulden, M.L., Lyons, E., Neff, J.C., Schuur, E.A.G., Zender, C.S., 2006. The impact of boreal forest fire on climate warming. Science 314, 1130–1132. Robitaille, A., Saucier, J.P., 1998. Paysages régionaux du Québec méridional. Les Publications du Québec, Sainte-Foy. Rotenberg, E., Yakir, D., 2010. Contribution of semi-arid forests to the climate system. Science 327, 451–454. Saucier, J.P., Grondin, P., Robitaille, A., Gosselin, J., Morneau, C., Richard, P.J.H., Brisson, J., Sirois, L., Leduc, A., Morin, H., Thiffault, É., Gauthier, S., Lavoie, C., Payette, S., 2009. Écologie forestière. In: Ordre des ingénieurs forestiers du Québec (Ed.), Manuel de foresterie. , 2e éd. Éditions Multimondes, Québec, pp. 165–315. Weaver, A.J., Eby, M., Wiebe, E.C., Bitz, C.M., Duffy, P.B., Ewen, T.L., Fanning, A.F., Holland, M.M., MacFadyen, A., Matthews, H.D., Meissner, K.J., Saenko, O.A., Schmittner, A., Wang, Z., Yoshimori, M., 2001. The UVic Earth System Climate Model: model description, climatology, and applications to past, present and future climates. Atmos. Ocean 39 (4), 361–428. Wolfe, R.E., Nishihama, M., Fleig, A.J., Kuyper, J.A., Roy, D.P., Storey, J.C., Patt, F.S., 2002. Achieving sub-pixel geolocation accuracy in support of MODIS land science. Remote Sens. Environ. 83, 31–49. Wulder, M.A., White, J.C., Cranny, M., Hall, R.J., Luther, J.E., Beaudoin, A., Goodenough, D.G., Dechka, J.A., 2008. Monitoring Canada’s forests. Part 1: Completion of the EOSD land cover project. Can. J. Remote Sens. 34, 549–562. Wulder, M.A., Franklin, S.E., White, J.C., Linke, J., Magnussen, S., 2006. An accuracy assessment framework for large-area land cover classification products derived from medium-resolution satellite data. Int. J. Remote Sens. 27, 663–683. Yakir, D., Rotenberg, E., 2010. Response. Science 328, 1480.
Contents lists available at ScienceDirect
Agricultural and Forest Meteorology journal homepage: www.elsevier.com/locate/agrformet
Boreal lichen woodlands: A possible negative feedback to climate change in eastern North America P.Y. Bernier a,∗ , R.L. Desjardins b , Y. Karimi-Zindashty b , D. Worth b , A. Beaudoin a , Y. Luo c , S. Wang d a
Canadian Forest Service, Natural Resources Canada, Québec, QC, Canada Agriculture and Agrifood Canada, Ottawa, ON, Canada Meteorological Service of Canada, Environment Canada, Ottawa, ON, Canada d Canada Centre for Remote Sensing, Natural Resources Canada, Ottawa, ON, Canada b c
a r t i c l e
i n f o
Article history: Received 24 September 2010 Received in revised form 15 November 2010 Accepted 25 December 2010 Keywords: Albedo Carbon sequestration Reforestation Fire regime
a b s t r a c t Because of successive forest fires, closed-canopy black spruce forests are susceptible to a shift towards open lichen–spruce woodlands in parts of the boreal forest of eastern North America. The shift from dark black spruce canopies to pale lichen ground cover offers a dramatic contrast in reflectance that may compensate for the CO2 emissions from forest fires in terms of radiative forcing. We have therefore looked at the climate change feedback that would result from the generation of lichen woodlands through changes in albedo and in stored carbon. Using albedo estimates based on MODIS imagery and incoming solar radiation for the period between 2000 and 2008 along with forest biomass estimates for eastern Canada, we have estimated that net radiative forcing for the conversion from closed-canopy coniferous forests to open lichen woodlands would be about −0.12 nW m−2 ha−1 , and would therefore generate a cooling effect in the atmosphere. Based on current estimates of area in open lichen woodlands within the closed-canopy black spruce–moss forests of eastern Canada, we estimate that a current net forcing of −0.094 mW m−2 has already arisen from such conversions. As projections of future climate have been linked to increased probability of forest fires, the generation of open lichen woodlands provides a possible negative feedback to climate change. Results also suggest that carbon sequestration through the afforestation of boreal lichen woodlands may not provide a climate change mitigation benefit. Crown Copyright © 2011 Published by Elsevier B.V. All rights reserved.
1. Introduction The boreal forest is Canada’s largest forest ecosystem, occupying 35% of the total Canadian land area and 77% of Canada’s total forest land (Natural Resources Canada, 2010). Fire regularly disturbs the boreal forest in Canada, leaving behind a temporary landscape of burnt trees and blackened earth. On average, about 9000 forest fires consume 2.1 million hectares annually in this ecosystem. While destructive, forest fires are a necessary and natural aspect of the life cycle of the boreal forest, releasing nutrients and rejuvenating the forest. The dominant tree species of the boreal forest are adapted and can regenerate following fires. Aspen (Populus tremuloides) regenerates vegetatively and abundantly from its root system following a fire, while black spruce (Picea mariana) and jack pine (Pinus banksiana Lamb.) have serotinous or semi-serotinous cones and heat-tolerant seeds that ensure a rapid re-seeding of the
∗ Corresponding author at: Canadian Forest Service, Natural Resources Canada, PO Box 10380, Stn. Ste-Foy, Québec, QC, Canada G1V 4C7. E-mail address: [email protected] (P.Y. Bernier).
ground where these species are present (Elliott-Fisk, 2000). Black spruce is a slow-growing species that can take 25 years to reach sexual maturity (Morneau and Payette, 1989). In eastern Canada, where pure stands of this species dominate the high-latitude boreal forest, a first burn or harvest followed by a second burn before trees reach sexual maturity may therefore eliminate or strongly reduce the tree cover, depending on the fire intensity. In welldrained landscapes of coarse glacial deposits, the now open ground often becomes covered with various species of lichens, mostly of the Cladina and Cladonia genus, creating stable open lichen woodland ecosystems that persist over time (Girard et al., 2008). Each land cover type has its own surface reflectivity, or albedo. When snow free, the boreal forest has an albedo of 9–11%, which tends to be less than that of other forest types and grasslands, which usually is in the 10–20% range. This difference in albedo is accentuated in winter, as snow blankets landscapes with sparse forest cover, resulting in an albedo in the range of up to 78%, but will tend to fall through a densely forested landscape, leaving the absorptive branches exposed and resulting in an albedo in the range of 10–26% (Betts, 2000; Betts et al., 2007a). Therefore, a change in land cover type has a direct effect on albedo, which in turn influences the local
0168-1923/$ – see front matter. Crown Copyright © 2011 Published by Elsevier B.V. All rights reserved. doi:10.1016/j.agrformet.2010.12.013
522
P.Y. Bernier et al. / Agricultural and Forest Meteorology 151 (2011) 521–528
radiation budget. This change can be expressed in terms of radiative forcing, which is defined as the disturbance of net irradiance at the tropopause after allowing the stratospheric temperature to readjust to radiative equilibrium (Chapter 6 in IPCC, 2001). Radiative forcing is often used to combine the effects of different agents, such as greenhouse gases, with a change in albedo into one metric that describes the net impact on climate. For instance, Betts (2000) showed that the radiative forcing of carbon sequestration and albedo change associated with afforestation/reforestation could be compared, and Pielke et al. (2002) estimated that for the boreal forest of North America this forcing ranged from 0.05 to 0.2 nW m−2 ha−1 . At the global level, a nominal climate sensitivity of 0.75 ◦ C per W m−2 of radiative forcing on a global scale has been reported by Hansen et al. (2008). Air temperature is increasing within the boreal region of Canada (Jones and Moberg, 2003), which may lead to an increase in the future area of forest burnt annually (Flannigan et al., 2005). In the boreal forest of eastern Canada, the burn rate since 1940 has decreased compared with historical rates (Le Goff et al., 2007), but predictions for a 2 × CO2 climate scenario suggest a 38% increase in future burn rates compared with current rates across the coniferous forest of Québec (Bergeron et al., 2006). It has been speculated that a climate change-induced increase in forest fires could lead to a positive climate feedback as carbon dioxide (CO2 ), methane (CH4 ) and nitrous oxide (N2 O) released during combustion promote further warming and burning (Kurz et al., 1995; Metsaranta et al., 2010). However, the larger impact may be through the change in surface albedo following a burn, particularly in early spring when solar radiation is increasing and snow still blankets recently burnt areas, whereas in forested areas most of the snowfall passes through or falls off the canopy, leaving the more absorptive branches exposed (Betts and Ball, 1997). Such a difference in albedo may result in a net cooling during winter (Bonan et al., 1992; Brovkin et al., 2004; Claussen et al., 2001; Ganopolski et al., 2001). Over the multiple decades required for forest regrowth, simulations suggest that the negative feedback of albedo could outweigh the positive feedback effects of CO2 emissions from the combustion of organic matter (Randerson et al., 2006). In the boreal forest of eastern Canada, the change in albedo following repeated fires may be particularly important because the ground cover of open lichen woodlands is very pale in colour with high albedo, and because these open systems are stable and not easily recolonised by trees. The increased rate of creation of open lichen woodlands through projected increases in fire frequency may therefore provide a negative feedback to climate change. The objectives of the present work were to evaluate the net radiative forcing resulting from the loss of biomass carbon to the atmosphere and the increase in albedo resulting from the creation of open lichen woodlands in the closed-canopy forest domain, and to assess the potential net radiative forcing effect of increased lichen woodland extent as a result of increased fire frequency. Results are also interpreted in terms of afforestation of open lichen woodlands.
2. Materials and methods 2.1. Study location The study area is located in the boreal forest of the province of Québec, Canada, and consists of the closed-canopy black spruce–moss bioclimatic domain (Robitaille and Saucier, 1998), located roughly between 49◦ N and 50◦ 30 N latitude, and covering an area of about 412,400 km2 (Saucier et al., 2009). This area is characterised by a low mean annual temperature (−2.5 ◦ C to 0.0 ◦ C), with annual precipitation increasing from west to east, from about 700–1200 mm. The dominant tree species is P. mariana (black
Fig. 1. Location of the reference blocks selected for the local dominance of one of four target cover densities of black spruce–lichen stand types. The darker shaded area corresponds to the black spruce–moss bioclimatic domain of the Quebec forest ecological classification of Robitaille and Saucier (1998).
spruce) with a variable abundance of P. banksiana (jack pine) and P. tremuloides (quaking aspen) in the dryer western portion of the domain, and of Abies balsamea (balsam fir) and Betula papyrifera (paper birch) in the wetter eastern portion of the domain. With the notable exception of the James Bay lowlands, most of these landscapes are underlain by coarse glacial deposits draped over hilly terrain. The historical fire return interval for this region varies across the landscape from about 140 years in the west to 280 years or more in the east. The tendency since 1940 has been a decrease in the fire frequency, with a doubling of the fire return interval to nearly 600 years in the east (Bergeron et al., 2006; Le Goff et al., 2007). Simulations done for 2 × CO2 and 3 × CO2 scenarios suggest a return to historical burn rates and possibly an increase in frequency above the historical rates for the western portions of the domain, with return intervals of 80 years for the 3 × CO2 scenario (Flannigan et al., 2005). Lichen woodlands are scattered in portions of this bioclimatic domain and have been expanding into the closed-canopy boreal forest for the past 1500 years. A possible cause of this expansion could be the drier late-Holocene conditions for this region (Jasinski and Payette, 2005), although fire risk seems to have been declining in the eastern boreal forest over the past 7000 years (Hély et al., 2010). 2.2. Selection of reference blocks Albedo values were determined for four forest types with a lichen ground cover, and with different canopy cover densities: 0–10%, 10–25%, 25–40% and 40–60%. The computation of albedo started with the selection of reference blocks for each cover density on a forest cover map that was developed as part of the Earth Observation for Sustainable Development (EOSD) project led by the Canadian Forest Service and funded by the Canadian Space Agency (Wulder et al., 2008). Location of the reference blocks is shown in Fig. 1. The circa 2000 land cover classification of Canada was completed in 2007 based on Landsat TM and ETM multispectral scenes. Within this national framework, the Enhancement-Classification Method (ECM) (Beaubien et al., 1999) was applied over the province of Québec using three Landsat spectral bands (red, NIR and MIR) from an assembly of 100 Landsat scenes acquired between 1998 and 2003. ECM was used to produce a detailed land cover map of Québec at a 25-m resolution with 52 thematic classes. These classes were further aggregated into the 23 EOSD thematic classes to be compliant with the national-level EOSD legend (Wulder et al., 2008). In our study, we used the detailed version of the land cover map of Québec to depict the forest covers of interest, i.e.
P.Y. Bernier et al. / Agricultural and Forest Meteorology 151 (2011) 521–528
523
Table 1 Number and area of EOSD-based reference blocks of lichen woodlands (EOSD ID 19 and 31) and of black spruce–lichen forests (EOSD ID 29 and 28) used for delineating MODIS pixels for albedo estimations, as well as the relative importance of these forest types within the 412,400 km2 black spruce–moss bioclimatic domain. Also shown is the number of field plots that were obtained for each forest cover density class from the EOSD plot database, and the mean values (with standard deviation in parenthesis) of total tree biomass in these plots. EOSD ID
19 31 29 28
Cover density class (%)
0–10 10–25 25–40 40–60
Number of reference blocks
7 5 4 4
Total area sampled (km2 )
17 32 3 77
conifer-dominated (75%+ of conifer species, mostly black spruce) with a lichen-dominated ground cover, and representing forest cover density classes of 0–10%, 10–25%, 25–40% and 40–60% (Table 1). Based on a regional accuracy assessment exercise conducted in the James Bay area in 2006 of the national-level EOSD land cover product and using a similar protocol as Wulder et al. (2006), overall classification accuracy was 73%, whereas specific class accuracy was 92%, 66% and 72%, respectively, for the 0–10%, the combined 20–25% and 25–40%, and the 40–60% density classes (without taking into account the different types of moss cover). More than 97% of the misclassified pixels were found in the immediate-neighbour density classes. Based on these accuracy figures, we can reasonably state that the accuracy for the four classes in this study is at least better than 75% considering the distinctive spectral signatures of black spruce stands with highly reflective lichen ground cover. For each cover density class of the black spruce–lichen forest type, we identified and extracted from four to seven small representative landscape blocks distributed across the black spruce–moss domain of Québec (Table 1). Each block had to have a minimum size of 0.5 km2 in order to include at least eight 250-m MODIS pixels, and had to be dominated by the forest cover type of interest. These blocks provided the location of nearly pure stands of the forest covers of interest for multi-spectral determination of albedo from the 250-m resolution MODIS images. The registration error for the 250-m MODIS pixel is usually less than 50 m (Wolfe et al., 2002), assuring a good spatial match between selected EOSD forest cover blocks and the MODIS pixels. Forest canopies present a continuum of cover density. However, experience in the photo-interpretation of forest cover shows that below a certain threshold tree cover ceases to be continuous and adopts a clumped structure. In commercial forestry, this threshold is set at 25% of the cover density class, where 25% is the lower bound of the last cover density class deemed to be potentially harvestable. Cover densities below 25% are simply classified as open woodlands. In the current study, we chose to adopt the same threshold between closed-canopy forests and open lichen woodlands and consider the two cover density classes of 0–10% and 10–25% to be lichen woodlands, while the two other denser classes are considered to be closed-canopy forests. Another element of comparison is the lichen-dominated ground cover. The most representative forest type in the black spruce–moss bioclimatic domain is that of pure black spruce stands with ground cover mosses such as Pleurozium schreberi (BSG) and Hylocomnium splendens (Hedw.). Mosses are dark green and will confer a lower albedo to the stand than white or grey lichen ground species such as Cladonia spp. and Cladina spp. For the present exercise, we chose to look only at the forest types with a lichen ground cover. By doing so, we consider only the change in albedo induced by the change in forest cover density and can therefore have a more direct comparison with radiative forcing associated with the carbon sequestered in trees.
Portion of bioclimatic domain (%)
1.9 3.1 4.6 1.6
Biomass
Number of plots
Mean (sd) (t ha−1 )
11 54 41 24
18.0 (9.8) 42.4 (19.5) 61.4 (21.8) 102.4 (30.5)
2.3. Solar radiation data Downward shortwave radiation was measured as a 30-min average at the Québec Flux Station–Eastern Old Black Spruce (49.693◦ N, 74.342◦ W), a CO2 flux measurement site of the Fluxnet Canada/Canadian Carbon Program (Margolis et al., 2006). This station has been operating continuously since 2003. Solar radiation is measured using a Kipp and Zonen CNR1 net radiometer, which measures the four component fluxes of net radiation. In the shortwave spectrum, the CNR1 integrates over the spectral range of 305–2800 nm. The CNR1 was mounted on a horizontal boom at a height of 22 m, approximately 8 m above a moderate density black spruce-dominated canopy with an understory of mosses and ericaceous shrubs (see Bergeron et al., 2007 for additional description of the site). The 30-min solar radiation values were used to produce monthly values of daily average incoming solar radiation over the 2003–2008 data record. As MODIS data were available for albedo determination for the longer 2000–2008 period, we used the 2003–2008 mean monthly values of daily solar radiation to match the 2000–2002 albedo data. Similar radiation measurements were also made at a companion site where trees had been harvested in 2000. Both radiation records were used to compute albedo, referred to below as “field albedo”, to be used for verification of MODIS albedo estimates. 2.4. Albedo data Surface albedo values were estimated using the Moderate Resolution Imaging Spectroradiometer (MODIS) located on board the TERRA satellite, available since 2000. The MODIS instrument provides global coverage every 1–2 days in 36 spectral bands ranging from visible to infrared and to thermal wavelengths between 405 and 14,385 nm. For the current exercise, in terms of albedo, we used the first seven spectral bands (B1–B7) with wavelength coverage of 459–2155 nm, which are specifically designed for land shortwave radiative applications. Two of them, B1 (620–670 nm) and B2 (841–876 nm), acquire imagery at a 250 m spatial resolution at nadir, while the rest (B3–B7) have a 500-m resolution. Bands B3–B7 were therefore downscaled to a 250-m resolution using a regression and normalization scheme for compatibility with B1 and B2. We first used bands B1 and B2 as inputs to generate the surrogate intermediate 250-m resolution images for each band from B3 to B7. The values in these images were then normalized to the observed values from the 500-m imagery in order to preserve the radiometric properties of the original data. Clouds and cloud shadows were detected and classified at a 250-m resolution by investigating multi-spectral features. The 10-day clear-sky composite data were then projected onto a standard LCC (Lambert Conformal Conic) map (Luo et al., 2008). A 10-day long interval is a practical compromise for MODIS to likely obtain clear-sky observations and to quickly capture any change in surface properties. The clear-sky composites of B1–B7 reflectance were then used to produce spatially
524
P.Y. Bernier et al. / Agricultural and Forest Meteorology 151 (2011) 521–528
continuous spectral albedo as in Luo et al. (2005). In that method, albedo is computed using a BRDF (Bi-directional Reflectance Distribution Function) model adjusted to observed reflectance from various angles but from the same land cover type. The modified Ross-Thick Li-Sparse BRDF model was chosen for its good performance in MODIS albedo retrieval (Luo et al., 2005). The seven spectral albedo data sets were then converted into one broadband albedo data set using the MODIS polynomial approximation conversion of Liang et al. (1999). This albedo value represents the overall solar short-wave albedo. The reference blocks selected from the EOSD land cover map for the four target forest types were remapped onto the LCC projection map. Their 25-m pixels were then aggregated into 250-m resolution pixels that matched the MODIS 250-m resolution pixels. Each aggregated pixel was assigned to the forest type that was dominant within the 25-m pixels used in the aggregation. For each block, pixels of the four forest types were grouped according to canopy cover densities and the albedo values from each group were averaged. The exact number of pixels for a given group within a given reference block varied over time because of cloud contamination or failure in albedo retrieval. Averaging was done only when there were at least four pixels available. For each month we obtained three albedo values from the three 10-day composite intervals, and therefore 27 values per month for the 2000–2008 period of record from which we computed the monthly albedo means and standard deviations. Multi-year monthly mean (and standard deviations) values of albedo per forest cover type were computed from the 27 monthly values obtained for the 2000–2008 record.
2.5. Forest carbon data
shortwave radiation in W m−2 (a) was estimated as: a = I(1 − ˛) + ˇ(˛I)
where I is measured downward shortwave radiation (W m−2 ), ˛ is albedo, and ˇ is a coefficient to correct for the atmospheric absorption of reflected shortwave radiation (Montenegro et al., 2009; Ramanathan et al., 1987; Weaver et al., 2001). The value of ˇ originally set to 0.3 by Weaver et al. (2001) was lowered to 0.23 by Montenegro et al. (2009) to account for atmospheric absorption of both downward and upward shortwave radiation. The net reflected shortwave radiation escaping the atmosphere is the result of the difference between incoming and absorbed radiation, which can be calculated as follows: NRS = I − a
(2) (W m−2 ).
where NRS is the net reflected shortwave radiation The differences in reflected shortwave radiation for various land cover types affect the local radiation budget. This effect on global radiative flux can be estimated for a given region of interest by: ıR =
(NRS1 − NRS2 )Ar Ae
2.6. Radiative forcing We calculated the radiative forcing attributable to change in albedo for each 250 m resolution pixel for the different forest densities as in Montenegro et al. (2009) and Betts (2000). Absorbed
(3)
where Ar is the area of the region of interest, Ae is the Earth’s surface area, and subscripts 1 and 2 refer to the different land cover types. Using this equation, the radiative forcing attributable to different land cover types can be determined. Biomass carbon contents obtained from the field plots of different forest cover densities were used to compute RFC , the equivalent radiative forcing, in nW m−2 ha−1 if this carbon was released into the atmosphere (Betts, 2000; Montenegro et al., 2009):
RFC = 5.35 ln 1 + Forest inventories had been carried out across the northern portion of the boreal forest for the explicit purpose of relating forest biomass to multi-spectral imagery and EOSD land cover classes in order to map the biomass of non-inventoried boreal forests (Beaudoin et al., 2007). Areas sampled (and average longitude and latitude of the plot locations) were Chibougamau (49.69◦ N, 74.34◦ W), Mistassini (50.90◦ N, 74.73◦ W), Radisson (55.55◦ N, 76.40◦ W), and Wabush (52.92◦ N, 66.99◦ W). Additional plots were measured in the James Bay lowlands of the black spruce–moss bioclimatic domain, but the mostly clay deposits and flat landscapes of these areas produce stands that follow different developmental dynamics than those from other portions of the domain. These plots were therefore left out. A total of 144 plots were retained for the analysis, with observed cover density values ranging from 5% to 85%. The tally of tree diameter at breast height and species in each plot was used to compute total aboveground biomass using the species-specific allometric equations of Lambert et al. (2005). The coarse root fraction was estimated at 18% of total biomass (Czapowskyj et al., 1985). Carbon concentration in the biomass was assumed to be 50%. Carbon in the non-biomass pools was left out in the present analysis as it was assumed that non-biomass carbon would be constant across the range of forest cover densities in the black spruce–lichen forest types. This assumption is supported by soil carbon data from open-canopy and closed-canopy black spruce and jack pine stands with lichen ground cover (Lagacé-Banville, 2009).
(1)
C C0
(4)
where C0 is the atmospheric CO2 concentration, and C is the change in atmospheric CO2 concentration if the carbon in the forest biomass is added to the atmosphere. Present day atmospheric CO2 concentration was assumed to be 385 ppmv (Menon et al., 2010). A value of RFC was thus obtained for the four cover density classes of black spruce–lichen forest types considered in the present study. 3. Results and discussion 3.1. Radiation and albedo data Downward solar radiation data show a strong seasonality typical of high-latitude locations (Fig. 2), with multi-year monthly average maximum and minimum daily downward shortwave radiation of 223 and 36 W m−2 recorded in June and December, respectively. Monthly average albedo values determined from MODIS observations for different forest densities (Fig. 3) show the expected increase in albedo with decreasing forest cover density. Albedo values are also highest when snow is on the ground, from December to April, with a maximum in February, and are lowest in summer. Albedo values start decreasing in late winter before the appearance of any bare ground resulting from the dirtying of the surface following initial melting events (Langham, 1981) and the faster discharge of snow from branches. Snow also enhances the difference in albedo among the different forest cover densities (Fig. 3). Snow-free monthly average albedo values of 0.16, 0.13, 0.11 and 0.09 for areas with less than 10%, 10–25%, 25–40% and 40–60% forest cover densities, respectively, are comparable to other published values. Betts and Ball (1997) report average summer albedo values of 0.2, 0.15 and 0.083 for areas covered by grass, aspen and conifer, respectively. Betts (2000) reports values of 0.18–0.21 for arable crop land and of 0.14–0.15 for dense evergreen coniferous forest under snow-free conditions. Furthermore, for deep snow
P.Y. Bernier et al. / Agricultural and Forest Meteorology 151 (2011) 521–528
525
Fig. 2. Multi-year monthly average of downward shortwave radiation for the 2003–2008 period at Fluxnet Canada/Canadian Carbon Program’s Québec Flux Station–Eastern Boreal Black Spruce. The error bar represents one standard deviation.
conditions, he reports a maximum albedo value of 0.26 for densely forested areas, which is somewhat lower than our winter maximum of 0.31 for the 40–60% canopy cover forest type. Monthly mean field albedo measurements obtained at the forested and harvested sites of the Québec Flux station bracketed the albedo estimates from the four target forest types, except in late spring to late summer when the albedo value of the harvested site was lower or equal to that of the lichen woodland with the lowest cover density (0–10%) (results not shown). 3.2. Radiative forcing Radiative forcing as a result of differences in albedo was computed as the difference in reflected shortwave radiation (Eq. (1)) between our highest forest cover density (40–60%) and each of the three other forest densities. Negative values reflect a reduction in absorbed solar radiation, and hence a cooling effect. Mean annual radiative forcing values of −0.26, −0.12 and −0.06 nW m−2 ha−1
were estimated for the forest cover types of 0–10%, 10–25% and 25–40% in comparison with 40–60% coverage, respectively (Fig. 4). The greatest radiative forcing for all cover densities occurred during the snow-covered season of January–April and peaked in mid-April (Fig. 4) because of the combination of pre-melting highalbedo snow (Fig. 5) and increasing shortwave radiation (Fig. 2). As expected, the most negative value of shortwave radiative forcing was obtained for the 0–10% cover density. This estimated albedo effect is not quite as pronounced as the 0.35 nW m−2 ha−1 calculated by Betts et al. (2007b) for albedo change resulting from afforestation, likely on account of differences in forest cover properties between the studies. Radiative forcing as a result of CO2 emissions following fire-driven loss of biomass was computed as the difference in RFC , the warming potential of the mean standing biomass if released into the atmosphere (Eq. (4)), between the 40–60% forest cover density type and the three lower cover density types. A positive value indicates a warming potential for the atmosphere. Mean annual values of radiative forcing of 0.13, 0.10
Fig. 3. Multi-year monthly averages of albedo determined for four forest cover types of different cover densities, with lichen ground cover, for the 2000–2008 period. The error bars represent one standard deviation.
526
P.Y. Bernier et al. / Agricultural and Forest Meteorology 151 (2011) 521–528
Fig. 4. Multi-year monthly average of radiative forcing effect of conversion from closed-canopy coniferous forest (40–60% cover density) to lower density closed-canopy forest and to open lichen woodlands for the 2000–2008 period. The error bar represents one standard deviation.
and 0.080 nW m−2 ha−1 were estimated for the forest cover types of 0–10%, 10–25% and 25–40% in comparison to the 40–60% forest cover type, respectively (Fig. 6). The largest value represents the impact of a change from closed-canopy forest to open lichen woodland, as observed following successive burns in that boreal environment. The net radiative forcing linked to the fire-driven creation of lower density forest canopies in these spruce–lichen environments is then computed as the sum of the albedo- and biomass-derived radiative forcings. Results from this computation (Fig. 6) suggest a net cooling effect of −0.12 nW m−2 ha−1 when the cover density drops from the 40–60% class to the 0–10% canopy density cover type, and an approximately neutral effect for a reduction from the 40–60% to the 10–25% and 25–40% cover types. These results suggest that the creation of very open lichen woodlands through repeated burning of closed-canopy forests has a net cooling effect on the atmosphere. The corollary to these results is that reforestation of open lichen woodlands and the creation of closed-canopy
forests may have at best no effect on global radiative forcing, or at worse a negative impact of 0.12 nW m−2 ha−1 for the conversion of 0–10% to 40–60% cover density forests. Currently, according to the EOSD-based classification (Table 1), open lichen woodlands in the 0–10% cover density class cover 1.9% of the black spruce–moss bioclimatic domain in the province of Québec, which represents an area of 7835 km2 . Therefore, gradual conversion of closed-canopy black spruce forests to lichen woodlands has to date generated a radiative forcing of −0.094 mW m−2 . However, it is at the moment difficult to estimate the current rate of lichen woodland generation throughout this domain, let alone predict how climate change will affect it. Given the functional relationship between successive disturbances and the generation of open lichen woodlands, these results nevertheless suggest a potential for an increased negative feedback between climate change and the boreal forest. Other studies have also suggested a warming effect due to afforestation because of changes in albedo (e.g. Rotenberg and
Fig. 5. Multi-year monthly averages of snow depth (m) for the 2003–2008 period at Fluxnet Canada/Canadian Carbon Program’s Québec Flux Station–Eastern Boreal Black Spruce.
P.Y. Bernier et al. / Agricultural and Forest Meteorology 151 (2011) 521–528
527
Fig. 6. Net radiative forcing resulting from the conversion from closed-canopy coniferous forest (40–60% cover density) to lower density closed canopy forest and to open lichen woodlands for the 2000–2008 period. The error bar represents one standard deviation.
Yakir, 2010). Of particular interest for our study, Betts (2000) calculated an increasing net warming effect due to afforestation with increasing latitudes in the boreal forests of eastern Canada. However, the very coarse resolution results in mixtures of land covers in each pixel, including the high proportion of natural and man-made water bodies in the boreal forest of Canada, thus making it difficult to attribute effects to specific forest cover types. Montenegro et al. (2009), on the other hand, suggest that high latitude afforestation/reforestation always results in a net cooling effect. Their study, however, was confined to agricultural areas, with high latitude calculations encompassing mostly areas in Eurasia with milder and drier climates and less snow cover, as well as ground cover types that may have albedo values lower than that of lichen. Their results are therefore valid for afforestation of land currently under cultivation, but may not be applicable to our present example which compares closed-canopy black spruce forests and open lichen woodlands. Our study specifically targets forest types along a gradient of forest cover densities down to open lichen woodlands, and is limited to land areas. It should therefore offer a more accurate evaluation of changes in albedo following a specific fire-induced forest cover change in the boreal environment.
Although albedo and carbon release and sequestration are likely first-order effects of land cover changes on the climate system, the total effect comprises more complex second-order interactions. Basic hydrological research has shown that the removal of trees reduces evapotranspiration (Bosch and Hewlett, 1982) and thus latent heat transfer, which can lead regionally to a reduction in cloud cover and of top-of-atmosphere albedo (Betts et al., 2007a). Forests also emit more longwave radiation than cooler high-albedo surfaces, but this radiation is reabsorbed by the atmosphere to varying degrees depending on the exact wavelength of the radiation. Ideally, all such effects should be incorporated into a common metric to compare the impact of land cover changes on the climate system, but this is still a very challenging task (see Lee, 2010; Leu, 2010; Yakir and Rotenberg, 2010). The issue is partially one of scale, with case studies looking at ground-level effects of local land cover changes whereas total radiative forcing must consider the larger effects on the global climate system. In areas where land cover drives socio-economic activities, impacts of land cover changes on energy use by society must also be included in such calculations. This is clearly an area in need of additional work. Acknowledgements
4. Conclusion Based on the computation of albedo changes and the impact of carbon release or sequestration on the atmospheric greenhouse effect, we conclude that the conversion of closed-canopy boreal forests to lichen woodlands through repeated fires has a cooling effect on the atmosphere of about −0.12 nW m−2 ha−1 , and that conversion to date has generated a total radiative forcing of −0.094 mW m−2 . Since climate change may enhance the natural conversion rate through increased fire frequency combined with forest harvesting, the additional generation of lichen woodlands would produce a negative feedback to climate change. Conversely, the conversion of open lichen woodlands into closedcanopy forests may generate a warming effect to the atmosphere. Calculations done with other types of boreal open woodlands, such as those dominated by ericaceous shrubs, may reveal a positive effect of reforestation. These results are important in their own right, but they also emphasize the need to incorporate changes in albedo in the evaluation of activities related to land use, land use change and forestry for climate change mitigation.
The authors thank the Fluxnet Canada/Canadian Carbon program for providing high-quality radiation data, as well as Stephen Côté and Philippe Villemaire of the Canadian Forest Service, Natural Resources Canada, for extracting reference blocks from the EOSD land cover map. Thanks are also extended to Dr. Alan Betts and to one anonymous reviewer for their constructive comments during the review process. The MODIS data were acquired from the NASA Goddard Space Flight Center, Level 1, and the Atmosphere Archive and Distribution System (LAADS) (http://ladsweb.nascom.nasa.gov). Part of the funding for this study was provided by the Agriculture and Agri-Food Canada program on Sustainable AGricultural Environmental Systems (SAGES). References Beaubien, J., Cihlar, J., Simard, G., Latifovic, R., 1999. Land cover from multiple thematic mapper scenes using a new enhancement-classification methodology. J. Geophys. Res. 104 (D22), 27900–27920. Beaudoin, A., Guindon, L., Hall, R.J., Luther, J.E., Fournier, R.A., 2007. A strategy to map biomass of Canadian northern boreal forests using multi-sensor/multi-
528
P.Y. Bernier et al. / Agricultural and Forest Meteorology 151 (2011) 521–528
resolution imagery: approach, methods and preliminary results. In: Proceedings of FORESTSAT 2007 Forests and Remote Sensing: Methods and Operational Tools , November 5–7, 2007, Montpellier, France. Bergeron, O., Margolis, H.A., Black, T.A., Coursolle, C., Dunn, A.L., Barr, A.G., Wofsy, S.C., 2007. Comparison of carbon dioxide fluxes over three boreal black spruce forests in Canada. Global Change Biol. 13 (1), 89–107. Bergeron, Y., Cyr, D., Drever, C.R., Flannigan, M., Gauthier, S., Kneeshaw, D., Lauzon, È., Leduc, A., Le Goff, H., Lesieur, D., Logan, K., 2006. Past, current, and future fire frequencies in Quebec’s commercial forests: implications for the cumulative effects of harvesting and fire on age-class structure and natural disturbancebased management. Can. J. Forest Res. 36, 2737–2744, doi:10.1139/X06-177. Betts, R.A., 2000. Offset of the potential carbon sink from boreal forestation by decreases in surface albedo. Nature 408, 187–190. Betts, A.K., Ball, J.H., 1997. Albedo over the boreal forest. J. Geophys. Res. 102 (D24), 28901–28909. Betts, A.K., Desjardins, R.L., Worth, D., 2007a. Impacts of agriculture, forest and cloud feedback on the surface energy budget in BOREAS. Agric. Forest Meteorol. 142, 156–169. Betts, R.A., Falloon, P.D., Klein Goldewijk, K., Ramankutty, N., 2007b. Biogeophysical effects of land use on climate: model simulations of radiative forcing and largescale temperature change. Agric. Forest Meteorol. 142, 216–233. Bonan, G.B., Pollard, D., Thompson, S.L., 1992. Effects of boreal forest vegetation on global climate. Nature 359, 716–718. Bosch, J.M., Hewlett, J.D., 1982. A review of catchment experiments to determine the effect of vegetation changes on water yield and evapotranspiration. J. Hydrol. 55, 3–23. Brovkin, V., Sitch, S., Von Bloh, W., Claussen, M., Bauer, E., Cramer, W., 2004. Role of land cover changes for atmospheric CO2 increase and climate change during the last 150 years. Global Change Biol. 10, 1253–1266. Claussen, M., Brovkin, V., Ganopolski, A., 2001. Biogeophysical versus biogeochemical feedbacks of large-scale land cover change. Geophys. Res. Lett. 28, 1011–1014. Czapowskyj, M.M., Robison, D.J., Briggs, R.D., White, E.H., 1985. Component biomass equations for black spruce in Maine. USDA Forest Serv. Res. Paper NE-564. Elliott-Fisk, D.L., 2000. The taiga and boreal forest. In: Barbour, M.G., Dwight Billings, W. (Eds.), North American Terrestrial Vegetation. , second ed. Cambridge University Press, Cambridge, pp. 41–74. Flannigan, M.D., Logan, K.A., Amiro, B.D., Skinner, W.R., Stocks, B.J., 2005. Future area burned in Canada. Climatic Change 72, 1–16. Ganopolski, A., Petoukhov, V., Rahmstorf, S., Brovkin, V., Claussen, M., Eliseev, A., Kubatzki, C., 2001. CLIMBER-2: a climate system model of intermediate complexity. Part II: Model sensitivity. Clim. Dynam. 17, 735–751. Girard, F., Payette, S., Gagnon, R., 2008. Rapid expansion of lichen woodlands within the closed-crown boreal forest zone over the last 50 years caused by stand disturbances in eastern Canada. J. Biogeogr. 35, 529–537. Hansen, J., Sato, M., Kharecha, P., Beerling, D., Berner, R., Masson-Delmonte, V., Pagani, M., Raymo, M., Royer, D.L., Zachos, J.C., 2008. Target atmospheric CO2 : where should humanity aim? Open Atmos. Sci. J. 2, 217–231. Hély, C., Girardin, M.P., Ali, A.A., Carcaillet, C., Brewer, S., Bergeron, Y., 2010. Eastern boreal North American wildfire risk of the past 7000 years: a model-data comparison. Geophys. Res. Lett. 37, L14709, doi:10.1029/2010GL043706. IPCC (Intergovernmental Panel on Climate Change), 2001. Climate Change 2001: The Scientific Basis. Contribution of Working Group I to the Third Assessment Report of the Intergovernmental Panel on Climate Change. Cambridge University Press, Cambridge. Jasinski, J.P., Payette, S., 2005. The creation of alternative stable states in the southern boreal forest, Québec, Canada. Ecol. Monogr. 75, 561–583. Jones, P.D., Moberg, A., 2003. Hemispheric and large-scale surface air temperature variations: an extensive revision and an update to 2001. J. Climate 16, 206–223. Kurz, W.A., Apps, M.J., Stocks, B.J., Volney, W.J.A., 1995. Global climate change: disturbance regimes and biospheric feedbacks of temperate and boreal forests. In: Woodwell, G.M., Mackenzie, F.T. (Eds.), Biotic Feedbacks in the Global Climatic System: Will the Warming Feed the Warming? Oxford University Press, Oxford, pp. 119–133. Lagacé-Banville, J., 2009. Caractérisation des stocks de carbone de 5 types de formations végétales dans un secteur du bassin versant de la rivière Eastmain, Baie James. M.Sc. Thesis, Université du Québec à Montréal, 159 pp. Lambert, M.-C., Ung, C.-H., Raulier, F., 2005. Canadian national tree aboveground biomass equations. Can. J. Forest Res. 35, 1996–2018.
Langham, E.J., 1981. Physics and properties of snow cover. In: Gray, D.M., Male, D.H. (Eds.), Handbook of Snow: Principles, Processes Management and Use. Pergamon Press, Oxford. Le Goff, H., Flannigan, M.D., Bergeron, Y., Girardin, M.P., 2007. Historical fire regime shifts related to climate teleconnections in the Waswanipi area, central Quebec, Canada. Int. J. Wildland Fire 16 (5), 607–618. Lee, X., 2010. Forests and climate: a warming paradox. Science 328, 1479. Leu, S., 2010. Forests and climate: the search for specifics. Science 328, 1479–1480. Liang, S., Strahler, A.H., Walthall, C., 1999. Retrieval of land surface albedo from satellite observations: a simulation study. J. Appl. Meteorol. 38, 712–725. Luo, Y., Trishchenko, A.P., Khlopenkov, K.V., 2008. Developing clear-sky, cloud and cloud shadow mask for producing clear-sky composites at 250-meter spatial resolution for the seven MODIS land bands over Canada and North America. Remote Sens. Environ. 112, 4167–4185. Luo, Y., Trishchenko, A.P., Latifovic, R., Li, Z., 2005. Surface bidirectional reflectance and albedo properties derived using a land cover-based approach with Moderate Resolution Imaging Spectroradiometer observations. J. Geophys. Res. 110, D01106, doi:10.1029/2004JD004741. Margolis, H.A., Flanagan, L.B., Amiro, B.D., 2006. The Fluxnet-Canada Research Network: influence of climate and disturbance on carbon cycling in forests and peatlands. Agric. Forest Meteorol. 140, 1–5. Menon, S., Akbari, H., Mahanama, S., Sednev, I., Levinson, R., 2010. Radiative forcing and temperature response to changes in urban albedos and associated CO2 offsets. Environ. Res. Lett. 5, 014005, doi:10.1088/1748-9326/5/1/014005. Metsaranta, J.M., Kurz, W.A., Neilson, E.T., Stinson, G., 2010. Implications of future disturbance regimes on the carbon balance of Canada’s managed forest (2010–2100). Tellus B, doi:10.1111/j.1600-0889.2010.00487.x. Montenegro, A., Eby, M., Mu, Q., Mulligan, M., Weaver, A.J., Wiebe, E.C., Zhao, M., 2009. The net carbon drawdown of small scale afforestation from satellite observations. Global Planet. Change 69, 195–204. Morneau, C., Payette, S., 1989. Postfire lichen-spruce woodland recovery at the limit of the boreal forest in northern Quebec. Can. J. Bot. 67, 2770–2782. Natural Resources Canada, 2010. The Atlas of Canada. http://atlas.nrcan.gc.ca/site/english/index.html. Pielke Sr., R.A., Marland, G., Betts, R.A., Chase, T.N., Eastman, J.L., Niles, J.O., Niyogi, D.D.S., Running, S.W., 2002. The influence of land-use change and landscape dynamics on the climate system: relevance to climate-change policy beyond the radiative effect of greenhouse gases. Philos. Trans. R. Soc. A 360, 1705– 1719. Ramanathan, V., Callis, L., Cess, R., Hansen, J., Isaksen, I., Kuhn, W., Lacis, A., Luther, F., Mahlman, J., Reck, R., Schlesinger, M., 1987. Climate–chemical interactions and effects of changing atmospheric trace gases. Rev. Geophys. 25 (7), 1441–1482. Randerson, J.T., Liu, H., Flanner, M.G., Chambers, S.D., Jin, Y., Hess, P.G., Pfister, G., Mack, M.C., Treseder, K.K., Welp, L.R., Chapin, F.S., Harden, J.W., Goulden, M.L., Lyons, E., Neff, J.C., Schuur, E.A.G., Zender, C.S., 2006. The impact of boreal forest fire on climate warming. Science 314, 1130–1132. Robitaille, A., Saucier, J.P., 1998. Paysages régionaux du Québec méridional. Les Publications du Québec, Sainte-Foy. Rotenberg, E., Yakir, D., 2010. Contribution of semi-arid forests to the climate system. Science 327, 451–454. Saucier, J.P., Grondin, P., Robitaille, A., Gosselin, J., Morneau, C., Richard, P.J.H., Brisson, J., Sirois, L., Leduc, A., Morin, H., Thiffault, É., Gauthier, S., Lavoie, C., Payette, S., 2009. Écologie forestière. In: Ordre des ingénieurs forestiers du Québec (Ed.), Manuel de foresterie. , 2e éd. Éditions Multimondes, Québec, pp. 165–315. Weaver, A.J., Eby, M., Wiebe, E.C., Bitz, C.M., Duffy, P.B., Ewen, T.L., Fanning, A.F., Holland, M.M., MacFadyen, A., Matthews, H.D., Meissner, K.J., Saenko, O.A., Schmittner, A., Wang, Z., Yoshimori, M., 2001. The UVic Earth System Climate Model: model description, climatology, and applications to past, present and future climates. Atmos. Ocean 39 (4), 361–428. Wolfe, R.E., Nishihama, M., Fleig, A.J., Kuyper, J.A., Roy, D.P., Storey, J.C., Patt, F.S., 2002. Achieving sub-pixel geolocation accuracy in support of MODIS land science. Remote Sens. Environ. 83, 31–49. Wulder, M.A., White, J.C., Cranny, M., Hall, R.J., Luther, J.E., Beaudoin, A., Goodenough, D.G., Dechka, J.A., 2008. Monitoring Canada’s forests. Part 1: Completion of the EOSD land cover project. Can. J. Remote Sens. 34, 549–562. Wulder, M.A., Franklin, S.E., White, J.C., Linke, J., Magnussen, S., 2006. An accuracy assessment framework for large-area land cover classification products derived from medium-resolution satellite data. Int. J. Remote Sens. 27, 663–683. Yakir, D., Rotenberg, E., 2010. Response. Science 328, 1480.