188.8.131.52 Additional factors not included in the models contribute uncertainty
Fire, insects and extreme events are not well modelled. Both forest composition and production are shaped by fire frequency, size, intensity and seasonality. There is evidence of both regional increase and decrease in fire activity (Goldammer and Mutch, 2001; Podur et al., 2002; Bergeron et al., 2004; Girardin et al., 2004; Mouillot and Field, 2005), with some of the changes linked to climate change (Gillett et al., 2004; Westerling et al., 2006). Climate change will interact with fuel type, ignition source and topography in determining future damage risks to the forest industry, especially for paper and pulp operations; fire hazards will also pose health threats (see Chapter 8, Section 8.2) and affect landscape recreational value. There is an uncertainty associated with many studies of climate change and forest fires (Shugart et al., 2003; Lemmen and Warren, 2004); however, current modelling studies suggest that increased temperatures and longer growing seasons will elevate fire risk in connection with increased aridity (Williams et al., 2001; Flannigan et al., 2005; Schlyter et al., 2006). For example, Crozier and Dwyer (2006) indicated the possibility of a 10% increase in the seasonal severity of fire hazard over much of the United States under changed climate, while Flannigan et al. (2005) projected as much as 74-118% increase of the area burned in Canada by the end of the 21st century under a 3 * CO2 scenario. However, much of this fire increase is expected in inaccessible boreal forest regions, so the effects of climate-induced wildfires on timber production may be more modest.
For many forest types, forest health questions are of great concern, with pest and disease outbreaks as major sources of natural disturbance. The effects vary from defoliation and growth loss to timber damage to massive forest die backs; it is very likely that these natural disturbances will be altered by climate change and will have an impact on forestry (Alig et al., 2004). Warmer temperatures have already enhanced the opportunities for insect spread across the landscape (Carroll et al., 2004; Crozier and Dwyer, 2006). Climate change can shift the current boundaries of insects and pathogens and modify tree physiology and tree defence. Modelling of climate change impacts on insect and pathogen outbreaks remains limited.
The effects of climate extremes on commercial forestry are region-specific and include reduced access to forestland, increased costs for road and facility maintenance, direct damage to trees by wind, snow, frost or ice; indirect damage from higher risks of wildfires and insect outbreaks, effects of wetter winters and early thaws on logging, etc. For example, in January 2005 Hurricane Gudrun, with maximum gusts of 43 m/s, damaged more than 60 million m3 of timber in Sweden, reducing the country’s log trade deficit by 30% (UNECE, 2006). Higher direct and indirect risks could affect timber supplies, market prices and cost of insurance (DeWalle et al., 2003). Globally, model predictions mentioned in the SAR suggested extensive forest die back and composition change; however, some of these effects may be mitigated (Shugart et al., 2003) and changes in forest composition will likely occur gradually (Hanson and Weltzin, 2000).
Interaction between multiple disturbances is very important for understanding climate change impacts on forestry. Wind events can damage trees through branch breaking, crown loss, trunk breakage or complete stand destruction. The damage might increase for faster-growing forests. This damage can be further aggravated by increased damage from insect outbreaks and wildfires (Fleming et al., 2002; Nabuurs et al., 2002). Severe drought increases mortality and is often combined with insect and pathogen damage and wildfires. For example, a positive feedback between deforestation, forest fragmentation, wildfire and increased frequency of droughts appears to exist in the Amazon basin, so that a warmer and drier regional climate may trigger massive deforestation (Laurance and Williamson, 2001; Laurance et al., 2004; Nepstad et al., 2004). Few, if any, models can simulate these effects.