Can Our Forests Take the Heat? Increasing Tree Mortality Rates Across the Western U.S. Phillip van Mantgem, PhD

Phillip van Mantgem is a research ecologist with the U.S. Geological Survey Western Ecological Research Center. He received his doctorate in ecology from the University of California Davis. His research interests include Conservation Biology, Fire Ecology, Forest Ecology and Global Change Biology. He’s currently leading the USGS Fire Severity Trends Project, which is testing the idea that climate may affect forest fire severity independent of fire intensity.

What I’d like to talk about today isn’t massive forest dieback, which is mostly what we’ve been hearing about so far [in this symposium]. What I’m interested in is slow, subtle changes that have been occurring over the last couple of decades in Western forests. I don’t have to convince anybody in this room that climate change is here already. No matter what resource you’re looking at, whether it is animal movements, timing of flowering of different plants, or the diebacks that we’re talking about today, we’re seeing the effects of climate change already.

In the western United States, one of the biggest effects we’re seeing is in the hydrologic cycle¬ – what’s happening with our water, the big limiting resource here. Scientists have been finding that the snowpack has been decreasing over time, and we have less snow. More precipitation is falling as rain rather than snow, so it doesn’t stick around as long. Stream flow is peaking earlier and earlier, which is another indication that things are getting dryer over the summer. What does that mean for forests?

One of our recent findings is that background tree mortality rates are increasing in old forests in the western United States. Detecting this trend is fairly easy because we have good data. Attribution, or what’s causing this trend, is more difficult. I’m going to present evidence today that the increasing background mortality rate is due to warming temperatures.

What is forest mortality rate and why are we concerned? Just like a human population, if you know how fast trees are growing, how fast they’re reproducing and how fast they’re dying, you actually know a lot about their population. Again, what I’m talking about today isn’t large-scale die-off, rather it’s the background mortality rate, which is a slow, subtle but important process. It’s important for a couple of reasons. First of all, you can kill a tree a lot faster than you can grow a new one. If you start changing the forest mortality rate, you can change a forest relatively quickly. What I’m going to be showing today is changes in tree mortality rates from about half a percent to about one-and-a-half percent over a couple of decades. That doesn’t seem very big, but it’s a compounded rate. If those changes are maintained year after year after year, it has a big effect.

For example, my wife and I are now shopping for a home loan, and we would be very concerned if interest rates were to rise by 1 percent! Again, it’s a compounded rate.

The idea that demographic rates might be changing over time isn’t new. In tropical systems, researchers have noticed an increase in what’s called recruitment, or birth, of new trees and mortality rates overtime. They’re seeing changes in stand biomass and in species composition. In the tropics, they’re seeing an increase of lianas, which are creeping, vine-like species. So we ask the question: can we see similar changes in western forests? Here’s the data that we used.

We were able to find 76 plots out of all the western United States that had the kind of information we needed to ask this question. We focused on old forests, forests that were more than 200 years old. We don’t expect old forests to have highly dynamic populations, so any changes to these populations are likely caused by external forces. We followed these plots over 20 years, tracking the fates of individual trees to get carefully measured demographic rates. Figure 1 shows our finding in those areas. The red symbols indicate sites where mortality rates were increasing; blue shows areas of decreasing mortality. The size of the dots indicates the magnitude of change we observed. What we’re seeing is that 87 percent of our plots have increasing mortality rates. That’s unusual. In these old forests, you wouldn’t expect rapid, directional change. We expect them to be demographically boring.

Figure 1: Locations of the 76 forest plots in the western United States and southwestern British Colombia. Red and blue symbols indicate, respectively, plots with increasing or decreasing mortality rates. Symbol size corresponds to annual fractional change in mortality rate. Redrawn from van Mantgem et al. 2009.

Over the short span of only two decades, it may be normal to have some instances of increasing mortality rates, but not 87 percent of the time. It’s like flipping a coin 100 times and getting 87 heads … something is going on. The mortality rate is doubling over an 18-year period, and we observed a trend towards increasing mortality over the entire length of our observations.

If background mortality rates are increasing, what’s causing that trend? It was a surprising result, so we thought that maybe there’s something in our data that’s really driving the overall trend. Maybe there’s some sort of artifact in there. We split out our data by regions, elevations, tree sizes, major species groups and historic fire frequencies, places that you’d expect to see a big impact of fire suppression or not. Essentially, we are seeing the same trend no matter where we look. Increasing tree mortality rates is a solid, robust, trend. And that trend is correlated with climate change.

Figure 2: Annual tree mortality rates from 1983 to 2004 for 21 permanent forest plots in the Sierra Nevada, California. Average water deficit (red line), an index of drought, predicted changes in the mortality rate (Redrawn from van Mantgem & Stephenson, 2007).

We used weather station data interpolated with the PRISM model to our particular study sites and found that changes in tree mortality rates correlated really well with changes in either temperature or water deficit (see Figure 2). Climatic water deficit is a good measure of water stress in plants. In the Sierra Nevada, we have some higher resolution data showing that tree mortality rates track changes in water deficit over time. The correlation is really tight. Unfortunately, it’s just a correlation, so we can’t say for certain what’s really driving this trend. If we want to predict what’s going to happen in the future, we really need to understand those mechanisms of tree mortality. We don’t have that level of understanding yet.

So why should you care about increasing mortality rates? One reason is that forests hold a lot of carbon. If you compare forests to other plant communities, an estimated 82 percent of the world’s terrestrial carbon is held up in forests. That amount compares rather favorably to the amount of carbon that is in the atmosphere right now. If all of our forests suddenly dropped dead and evaporated, the level of atmospheric carbon would essentially double. I don’t think that’s going to happen, but our data suggest that our forests might be slowly leaking carbon.

Increasing background tree mortality rates also suggest that our forests are being subjected to chronic stress. When we get an acute stress like a drought event, they might be more susceptible to massive dieback events. Figure 3 is a photo taken by Craig Allen of a dieback event in the Four Corners region of the western U.S. We think that such dieback events might be more common in the future — not only as climate changes, but as forest populations become more and more susceptible to those stresses.

This type of long-term research is very important to help people make critical decisions about our forests. The work that I’ve been showing you today is boots-on-the-ground research. We can’t get this from remote sensing. We can’t model our way to these results. The only way to do it was to have people out there counting trees and measuring trees year after year after year. We need to be committed to getting these measurements and seeing what the patterns they reveal. There has not always been a great deal of support for monitoring programs. Everybody wants this kind of data, but it is important to acknowledge the need for funding to support for these studies.

If things are actually changing, then we are going to have to rethink how we manage our forests. The National Park Service, for instance takes the idea of “naturalness” very seriously. Whenever possible, they try to restore and maintain naturally functioning ecosystems. When that’s not possible, they work to maintain the closest approximation of that “natural condition,” using 1850 as a target period because it was a time prior to the significant changes to the landscape that occurred with mechanized development. It has been a good model and probably still is in many cases. However, with climate change, we no longer have environmental conditions similar to 1850.

My research suggests we are going to have to re-think how we’re going to manage these areas in this new context. It might mean trying some different management approaches. For example, do we want to use more intense prescribed fires to thin out some of our forests? Should we start engaging in assisted species migration? I don’t think we have the answers for that yet, but we need to start asking these sorts of questions. Once we start asking such questions, we hope that there will be some light at the end of the tunnel.

Figure 3: Die-back event of piñon pine in the American Southwest. Photo credit: Craig Allen.