Warming and drying climate, bark beetle outbreaks, and wildfire all pose challenges to western US conifer forests. Widespread bark beetle outbreaks have been impacting large swaths of western North America. In the Lake Tahoe Basin, there is a strong link between severe drought and beetle outbreaks. Beetle outbreaks can also be affected by the density of their hosts because many beetle species use specific tree genera. As an example, mountain pine beetle only uses pine trees for hosts.
In drier, more fire-prone forests, fire-exclusion has increased tree density and in some cases host density. This change in structure has also increased the risk of stand-replacing fire. Since management activities to reduce the risk of stand-replacing fire typically involve reducing tree density and restoring surface fire, we sought to determine if these treatments might also reduce beetle outbreak potential.
In a study of Lake Tahoe Basin forests, led by Rob Scheller, we used the LANDIS-II simulation model to determine if beetle outbreaks would increase with climate change and if management activities to reduce wildfire hazard would reduce the impacts of beetle outbreaks. We hypothesized that climate change would increase beetle outbreaks and reduce carbon uptake by the forest and that management activities would reduce beetle-caused mortality and increase carbon uptake by the forest.
We found that climate change without beetles caused a reduction in carbon stored in trees and that beetles without climate change caused a reduction in carbon stored in trees. However, the combined effects of climate change and bark beetles caused a large reduction in the amount of carbon stored in trees across the study area (see Figure 1).
White fir and Jeffrey pine are two of the more common species in these forests. Previous research has shown that prior to fire-exclusion, Jeffrey pine was more common than white fir and with fire exclusion, white fir has become more common. In areas around communities, managers often focus harvesting efforts on white fir to reduce fire hazard. When we looked at these individual species in the areas treated around communities, we found that without management bark beetles caused both species to decline (see Figure 2). When we simulated management and no bark beetles, white fir decreased and Jeffrey pine increased. The combined effects of beetles and management really reduce the amount of white fir and resulted in a similar amount of Jeffrey pine as the simulation that only included beetles.
Our management simulations did not include treating all forests within the Basin. Treatments focused on communities and roadways. Because climate projections for the Basin get warmer and drier later in the 21st century, we looked at the chance that these forests become a source of carbon to the atmosphere with continued climate change and bark beetles. We found that the chance of these forests become a source of carbon to the atmosphere increases later this century with both beetles alone and beetles and management combined. Reducing the impacts of beetles may require more thinning to further reduce host density. These results show that we have to consider the full suite of disturbance agents when trying to determine the best path forward for managing forests under climate change.
The carbon carrying capacity of an ecosystem is the maximum amount of carbon that can be sustained in a given location on the planet under prevailing climate and natural disturbance conditions. The idea is that given the temperature range and amount of precipitation at a given location, there are only certain types of plant species that can grow there. This is exactly why we don’t see giant sequoia trees growing next to Joshua trees in the Mojave desert.
rests in the Sierra are heavily dependent on the winter snowpack for providing growing season moisture. Under a moderate-high emission scenario, the snowline elevation is projected to increase by about 200 meters (650 feet) of elevation over the next 100 years. Now, as any geographer will tell you when you go up in elevation there is less land area. This increasing snowline elevation means there will be considerably less area that maintains a snowpack to provide water during the growing season.
In a recent paper led by Shuang Liang, we ran simulations of forests of the Sierra Nevada under projected climate change and wildfire. We used climate projects from three climate models and area burned projections from Leroy Westerling. We ran our simulations out over several hundred years and assumed that the climate beyond the year 2100 would be similar to the climate projected for 2090-2100. These long simulation times allowed us to evaluate the effects of altered climate on the carbon carrying capacity of currently forested areas in the Sierra Nevada.
Our simulations show that with increased warming and drying and more area burned by wildfire that the amount of carbon these ecosystems can sustain will decrease by as much as 73% (top panel). A big part of that reduction is due to as much as a 65% reduction in forested area (middle panel). The loss of forest cover is driven by fewer species of tree seedlings being able to establish in these future conditions, especially after more area is burned by wildfire.
As the figure shows, there is quite a bit of variability between the different climate model projections. The GFDL (red) and CNRM (green) models show a large decline in both carbon and forested area, whereas the CCSM3 (blue) climate model shows both carbon and forested area holding pretty steady. The loss of forest cover leads to the carbon flux results in the bottom panel. The bottom panel shows the net ecosystem carbon balance. When the values are positive, the forests are removing carbon from the atmosphere and when the values are negative, the forests are a source of carbon to the atmosphere. Under two of the climate model projections, Sierran forests become sources of carbon to the atmosphere and will be another thing we have to consider when thinking about mitigating climate change.
The big question is – what can we do about the prospect of having 50% of the forested area converting to something other than forest? Reducing our greenhouse gas emissions to the atmosphere is the logical place to start, but since this blog is about forest ecology research you will have to stay tuned for the results of our current work that is looking at the role of management in reducing the risk of forest cover loss to climate change and wildfire.
We were out on 18 April 2017 to conduct a post-snow melt, pre-dry season seedling survival survey on our plots in the Jemez Mountains. You can read more about the overall project here, but a quick recap:
We planted seedlings from four species stratified based on aspect (north, south) and cover (shrub, no-shrub) this past fall in the footprint of the 2011 Las Conchas Fire.
We found that our elk deterrent fencing was inadequate in a few areas and that mmmm, soil moisture and temperature sensor wires sure do look tasty to an elk. S.H. Hurlbert in his classic 1984 paper on pseudoreplication states that replication controls for, among other things, “non-demonic intrusion”. He then goes on to state “If you worked in areas inhabited by demons you would be in trouble regardless of the perfection of your experimental designs. If a demon chose to “do something” to each experimental unit in treatment A but to no experimental unit in treatment B, and if his/her/its visit went undetected, the results would be misleading.” Further in the same paragraph he states “Whether such non-malevolent entities are regarded as demons or whether on simply attributes the problem to the experimenter’s lack of foresight and inadequacy of procedural controls is a subjective matter.”
I guess in the case of our planted seedlings, we foresaw the potential for elk to eat our seedlings and therefore exhibited some experimental foresight. However, we didn’t foresee the appeal of sensor cables to elk. Either way, it looks like we’ve got a case of “non-demonic intrusion” and a case of “demonic intrusion”. Fortunately, it is pretty clear when an elk decides to either eat the tasty seedling treat or just ripped it out of the ground and we can look at the effect of including this information or excluding it in a summary presentation of the information.
Even though I still don’t have any resolution regarding whether or not elk qualify as demons, we can at least look at initial survival of the four different seedling species. The labels are as follows: PSME = Douglas-fir, PIPO = ponderosa pine, PIED = pinyon pine, and PIST = southwestern white pine. The strata are labeled as follows: NO = north aspect without shrubs, NS = north aspect with shrubs, SO = south aspect without shrubs, and SS = south aspect with shrubs.
We had higher percent survival on north aspects for all species and lower percent survival on south aspects. Southwestern white pine (PIST) seemed to have an especially tough time on south aspects, as compared to the other species.
One of our hypotheses for this project is that seedlings will have higher survival on north aspects than on south aspects and that on south aspects, survival will be higher with shrubs than without. We’ll see how this plays out during the May-June dry period. So, stay tuned for the next post where we’ll also have some temperature and relative humidity data by strata as well.
Postdoc position in ecosystem modeling
The Earth Systems Ecology Lab (www.hurteaulab.org) at the University of New Mexico is recruiting a postdoctoral researcher with a strong background in ecosystem modeling and programming to contribute to a project aimed at understanding the interaction of climate change and wildfire on post-fire forest recovery. This project will integrate tree seedling data, flux tower data, and ecosystem modeling with the objective of understanding how changing climate will alter forest recovery following wildfire.
The initial appointment is for one year (beginning summer 2017), with the possibility of extension for up to two additional years. A competitive salary and benefits will be provided. Required qualifications include a PhD in ecology, ecosystem science, earth/environmental sciences, or statistics and programming experience with R or Python and C+ or C#. Willingness to periodically participate in field sampling is desirable.
Applicants should submit a cover letter detailing research interests and goals, a complete CV, and names and contact information for three references to Matthew Hurteau (firstname.lastname@example.org). Review of applications will begin on 28 April 2017.
The University of New Mexico is committed to hiring and retaining a diverse workforce. We are an Equal Opportunity Employer, making decisions without regard to race, color, religion, sex, sexual orientation, gender identity, national origin, age, veteran status, disability, or any other protected class.
Shuang Liang successfully defended her dissertation entitled "Simulating the effects of climate change, wildfire, and fuel treatment on Sierra Nevada forests". Congratulations Shuang!
The fire triangle (oxygen, heat, fuel) applies to wildland fire as much as it does to any other type of fire. Fire suppression involves “removing” one of the legs of the triangle, usually fuel. The fire behavior triangle (fuels, topography, weather) influences how fire behaves when burning through a forest. Things like steep slopes, high temperatures, and gusty winds can all cause more active fire behavior. In a recent paper, Brandon Collins analyzed 40 years of weather data in the northern Sierra Nevada. He found that the frequency of extreme fire weather (hotter, drier, and windier conditions) has increased. We tend to get our largest and hottest wildfires under extreme fire weather.
In a recent paper led by Dan Krofcheck, we set out to better understand how extreme fire weather influences the effects of forest treatments on moderating fire behavior (i.e. surface fire rather than crown fire) in the Dinkey Creek watershed in the southern Sierra Nevada. Since we wanted to isolate the effects of fire weather, we created two different sets of fire weather inputs. The first was generated using 13 years of data from three different weather stations located in or near the Dinkey Creek watershed. The second set of weather was generated using weather station data from the 2013 Rim Fire, which burned 141,131 ha of the Stanislaus National Forest and Yosemite National Park. The figure below shows the differences in model inputs generated from these two different sets of weather data. The distributions from the 13 years of weather data (labeled contemporary) have a lower build-up index (panel B) and fine fuel moisture code (panel C), which are measures of how much moisture is in the fuel and lower values mean the fuel is wetter and less available to burn. While these two parameters both had fairly big differences, the biggest difference was in mean daily wind speed (panel D). The weather station near the Rim Fire (labeled extreme) recorded much higher daily wind speeds than we see in the contemporary data.
We ran simulations with no management activities, thinning, and thinning followed by regular prescribed burning under both distributions of weather. Under contemporary fire weather, we didn’t find a big difference between the management scenarios in terms of fire severity (0 = no fire; 5 = tree-killing fire). When we ran the same management scenarios with the extreme fire weather, the results were quite striking (Figure 2).
The no-management and thin-only scenarios both had large increases in fire severity across the majority of the watershed (Figure 3). We also found increasing fire severity under the thin and maintenance burning scenario, but across the majority of the watershed, severity was considerably lower with regular prescribed fire use. When we looked at the variability in fire severity between model runs for a particular management scenario we found that the no-management and thin-only scenarios consistently burned more severely.
The most interesting finding for us was the large difference between the thin-only and thin and maintenance burning treatments. The increase in fire severity for the thin-only treatment was driven by the big increase in shrub growth that occurred following thinning. Opening up the canopy allows more light to reach the forest floor and shrubs are able to grow in more continuous patches that carry fire. When we followed the thinning with regular prescribed fire, the repeated surface fire held the shrubs in check and reduced the fuel available to carry fire. As fire weather continues to become more extreme with changing climate, our results suggest that restoring surface fire is central to reducing the chance of high severity, tree-killing wildfire.
In a previous study we looked at the effects of forest treatments to restore surface fire and how they influence forest carbon storage. That study used historical climate to grow the forests, which left me wondering if how carbon dynamics and the influence of restoring surface fires might be altered by changing climate. To answer this question, I used the same forest landscape model as the previous study and obtained projected climate data from the climate models that were run in support of the IPCC’s Fifth Assessment Report.
Using all of these climate projections, I developed monthly climate distributions for early, middle, and late 21st century. As the figure on the right shows, each period is progressively warmer and the amount of precipitation falling each month is quite variable.
The figure below shows the influence of climate change and wildfire without treatments to restore surface fire (solid lines) and with treatments to restore surface fire (dotted lines). Without treatments, the combination of tree-killing wildfire and warmer temperatures cause the amount of carbon stored in ponderosa pine trees to decline over time. With treatments, the amount of carbon initially decreases because of the removal of trees by thinning and then increases over time. This is because fewer ponderosa pine trees are killed when wildfire occurs. The treatment simulations (dashed lines) also show how climate affects tree growth. The blue line and shaded area represents carbon under late-century climate (2090-2099) and it is significantly lower than under early-century climate (2010-2019).
Interestingly, the results from simulations that don’t include treatments are heavily influenced by a large increase in Gambel oak. This species can resprout after fire and provide a fuel source to carry a subsequent fire before the pine trees are large enough to withstand fire. As shown in the photos below, this is an outcome that we are already seeing in the southwestern US following large, hot wildfires.
The results from this study suggest that by restoring natural fires to this ponderosa pine forest, we can maintain a pine forest for a longer period of time under changing climate.
Trees can live for hundreds to thousands of years. With that kind of lifespan they certainly experience a range of climatic conditions. When we move from thinking about the tree to thinking about the forest and how changing climate might impact a forest, we’ve got to consider how climate influences tree regeneration. There is a body of research that has examined the climatic conditions under which different tree species will regenerate. The range of climatic conditions that a seedling will tolerate is generally much narrower than the range of conditions a mature tree will tolerate.
In a recent study led by Shuang Liang, we examined how future climate change and wildfire might alter the distribution of tree species across the Sierra Nevada Mountains of California and Nevada. We used the LANDIS-II simulation model, projected climate data from three climate models, and area burned projections to simulate the forests of the Sierra Nevada. The climate models show that with unabated human carbon emissions we can expect about 3-5°C of warming by 2099, which will dry out the environment leaving forests with less water for growth.
When we ran simulations we used the three climate projections in Figure 1 and we also ran simulations with climate from the period 1980-2010 to create the baseline scenario. This baseline case included area burned data from the same period and provides a comparison with the future climate and wildfire scenarios. When we looked at how future climate and wildfire impacted where different tree species were on the landscape, we found small changes for the mature trees (Fig 2 a,b). At the lowest elevations, we found slight declines in tree species like white fir that prefer more precipitation, but changes in the distributions of other species were quite small. However, when we looked at tree regeneration, we found large differences between the baseline scenario and the future climate scenario. In the middle elevation band (3900-6900 feet), we found sharp declines in the amount of regeneration events for the more moisture loving species like white fir and we found that more drought-tolerant species like ponderosa pine had accounted for more of the regeneration (Fig 2c,d).
Fig 2: The spatial distribution of dominant tree species by biomass (a) and by elevation band (b). The spatial distribution of dominant tree species by number of regeneration events (c) and by elevation band (d). The results use baseline climate and wildfire (BSWF) and projected climate and wildfire (CCWF).
We also found sharp declines in the number of regeneration events over the 90 year simulation (Fig 3). The majority of the mountain range had 50% fewer regeneration events with future climate and wildfire than did the baseline scenario.
When you’ve got forests that experience wildfire, reduced regeneration is important because it means that areas that burn will take longer to recover to forest. When we compared the percentage of the Sierra Nevada that was not forested in 2099, we found that it had increased by about 5% over the baseline climate scenario. Now, 5% doesn’t seem like much, but when you are talking about the whole Sierra Nevada mountain range that is approximately 170,000 hectares (656 square miles).
These results suggest that we can expect some pretty large changes in the species that make up the forests of the Sierra Nevada and the amount of area that has forest cover as the climate changes.
2016 has been an interesting year for wildfire research. A couple of studies published this year have identified linkages between the increasing area burned by wildfire and increasing temperature. Leroy Westerling published a study where he looked at the increase in area burned by large wildfires over the period 1970-2012. He found that across the western US, area burned by large wildfires has increased by 556% over the 1983-1992 average (Figure 1 top). His analysis shows that increasing temperatures correlate with longer fire seasons. Average fire season length increased by 84 days between the first decade of his analysis (1973-1982) and the last decade (2003-2012)
John Abatzoglou and Park Williams published a study where they looked at the relationship between fuel aridity and area burned in the western US. Fuel aridity is a measure of how dry the material in the forest is and the drier it is the more flammable it is. Their results show a strong relationship between this measure of dryness and area burned (Figure 1 bottom). Increasing temperature is also playing a role here and they attribute approximately half of the forest area burned by wildfire to human-caused climate change over the period 1984-2015.
Fire suppression costs by year from the National Interagency Fire Center website show that from 1985 to 2015 we spent approximately $36.6 billion on fire suppression in 2015 dollars. While the year with the highest total fire suppression cost was 2015 ($2.13 billion), the year with the highest per acre suppression cost was 1998 ($455/acre, Figure 2). Area burned in any particular year explains about 51% of the variability in suppression costs. A number of factors account for the remainder of the variability in suppression cost, including proximity to developed areas. As an example, the 2016 Soberanes fire in southern California burned 132,127 acres and cost an estimated $260 million to suppress, that works out to $1967/acre.
When we look at how decadal averages in suppression cost have change over time, the picture is similar to Leroy Westerling’s results (Figure 3). It is important to note that the first average suppression cost only covers the period 1985-1992 and the last average suppression cost bar is only for the years 2013-2015. The majority of the suppression expenditures are by the US Forest Service and in a 2015 report they showed that fire suppression accounted for 52% of their budget. Given that area burned by wildfire in the western US is increasing as the temperature goes up and suppression expenditures are on the rise, this really begs the question – how sustainable is our current relationship with fire?