Overstory Tree Mortality in Ponderosa Pine and Spruce-Fir Ecosystems Following a Drought in Northern New Mexico

Drought-caused tree dieback is an issue around the world as climates change and many areas become dryer and hotter. A drought from 1998–2004 resulted in a significant tree dieback event in many of the wooded areas in portions of the Jemez Mountains and the adjacent Pajarito Plateau in northern New Mexico. The objectives of this study were to evaluate and quantify the differences in tree mortality before and after a recent drought in ponderosa pine and spruce-fir ecosystems, and to assess the effect of mechanical thinning on ponderosa pine mortality. Significant increases in mortality were observed in the unthinned ponderosa pine ecosystem. Mortality varied significantly between species and within size classes. Mechanical thinning of ponderosa pines reduced overstory mortality to non-significant levels. A lack of rainfall, snowfall, and increases in daily minimum temperature contributed most to the mortality. Adaptive management, including the use of thinning activities, appear to moderate the impact of climate change on ponderosa pine forests in this region, increasing the long-term health of the ecosystem. The impact of climate change on the spruce-fir ecosystems may accelerate successional changes.


29
Prolonged below-normal precipitation and above-normal temperatures from 1998-2004 led to

38
Most tree death is episodic and irregular [7]; this is especially true in dieback cases. These 39 essentially episodic but unpredictable events appear to occur somewhat regularly throughout 40 geologic time. The causes of dieback are case by case due to the differing conditions of each area, 41 but long-term climatic change has been suggested as a major cause of stand-level dieback [8].

42
The invasion of woody species into areas they did not historically inhabit and the general 43 increase of tree density in areas already forested has created greater water stress in many areas of the American Southwest. This in turn could mean that the water stress caused by drought would be 45 exacerbated due to the excess amount of woody species. For ponderosa pine, the spatial pattern of 46 mortality corresponded directly to elevation/moisture gradients. Mortality of ponderosa pine was 47 apparently widespread on drier, lower-elevation sites across the drought-affected region.

48
Moreover, the effects of the drought have persisted. There is little evidence of ponderosa pine 49 re-establishment in recent decades in spite of favorable climatic conditions 9]. In Northern Arizona 50 and Northern New Mexico a more recent drought caused substantial tree dieback in many species, 51 with piñon pine being the most affected [1 10, 11]. The proximal cause of the mortality for most of 52 the trees was apparently infestation by bark beetles; such outbreaks are tightly tied to 53 drought-induced water stress [1]. It was found that piñon seedlings exhibited the lowest levels of

65
The Los Alamos region, located on the east slopes of the Jemez Mountains of northern New 66 Mexico, is relatively free of human-caused disturbance, contains a large elevational gradient of 67 habitats, has a history of strong environmental assessment, and has an abundance of data available      Significant increases in temperatures and decreases in total precipitation occurred during the 98 drought, even though no significant reductions in annual snow pack was found. 99

123
The subplots were numbered sequentially in a clockwise direction, with the subplot number 1 124 located in the upper left corner of the macroplot. In this fashion, subplot 1 and subplot 2 were 125 always upslope, whereas subplot 3 and subplot 4 were always downslope. Two of these subplots 126 were randomly selected for further sampling.

140
To optimize the sampling of shrubs and small trees that were less than ten feet tall, a system of 141 strip plots was selected. This was accomplished by assessing the density of shrubs and small trees 142 in the less than 3 m size height class and selecting an appropriate strip plot scheme. More compact 143 combinations of strip plots and sub-strip plots were used for plots that have numerous seedlings, 144 saplings and small shrubs. Moderate combinations were used for typical forested conditions.

145
Extensive combinations of strip plots and sub-strip plots were used for sparsely forested or 146 non-forested plots that had few seedlings, saplings or small shrubs. Using the optimal plot size that 147 was selected and recorded, two line transects were constructed in each quad, and also used to

156
Sampling occurred between 1998 and 2005 to capture a range of years, overlapping the drought 157 in question. All plots overstory densities ranged from 500-800 trees per hectare, depending on the 158 commutiy type and site conditions. Thinning targets ranged from 50-1500 trees per hectare, again 159 depending on which community was thinned. Thinned plots were within management areas to be 160 thinned, and occurred across the range of years. They were not set up as paired to non-thinned 161 plots, nor was proximity to unthinned plots measured.

165
Strip plots were used to sample all trees and shrubs that met specific height criteria. The strip plots 166 were used to sample all trees and shrubs above 0.6 m tall but less than 3 m in height. Information

207
Since no significant differences with respect to pre and post-drought measurements for thinned 208 ponderosa pine were found (Table 1), CCA analyses were not run. However, Wilcoxon 209 paired-sample one-tailed t tests found significant greater mortality post-drought (Table 2).

210
The only size classes that were found to be significantly different with respect to mortality pre 211 and post-drought in Spruce-fir were the DBH size classes of 20-30 and 30-40 cm (Table 1), with 212 greater mortality recorded post-drought. While no size classes were found to be significant for the 213 Abla/Vamy Habitat Type, theThe 10-20, 20-30, and 30-40 cm size classes were found to be significant 214 greater post-drought in the Pien/Erex Habitat Type (Table 3).  for Douglas-fir only the 0-10 cm class. No size classes for aspen that were found to be significantly 234 different (Table 4).

236
While total mortality of all trees in the Spruce-fir communities ranged from 2.3-5.4% across the 237 five species, the percent mortality of that which was drought induced ranged from 21.7 to 75.0 % for 238 the conifer species, but only 8.5% for aspen (Table 5).

241
Average percent mortality of all species (annual mortality) correlated to the annual (calendar

242
year) snow, rainfall, and average daily maximum and minimum temperatures showed that 243 unthinned ponderosa pine mortality was highly related to rainfall and snow (Figure 4). Douglas-fir 244 mortality was to a lesser extent further related to snow and rainfall, and white fir mortality was

256
Average percent mortality of all species was correlated to the annual water year snow, rainfall, 257 and number of heating and cooling degree-days ( Figure 5). Unthinned ponderosa pine mortality,

258
and Douglas-fir to a lesser extent, was highly related to rainfall and snow. White fir mortality was 259 highly related to increases in temperatures, and aspen showed a relationship to heating days.

260
Engelmann spruce showed a positive relationship to cooling days.

272
The three smallest size classes showed some relation to maximum temperature.

292
There were substantial differences in the pre-drought stands and the post-drought unthinned 293 ponderosa pine stands. The average percent mortality was approximately ten percent higher in the 294 post-drought stands with larger ponderosa pine trees more significantly impacted by the drought.

295
The drought had a significant impact on ponderosa pine mortality. The primary factors influencing 296 the mortality were decreases in precipitation and increases in temperature. There were no 297 significant differences in the pre-drought stands and the post-drought thinned ponderosa pine 298 stands. This is very different from the unthinned ponderosa pine data, suggesting that the thinning 299 process significantly reduced ponderosa pine mortality.

300
When analyzed by size class, irrespective of species, the data showed significant differences in 301 the 20-30 and 30-40 cm DBH size classes, but not for ponderosa pine. The main difference between 302 the two sets of data is that the spruce-fir data showed substantial mortality in the pre-drought years.

303
There are several potential reasons for this occurrence. The stands in the spruce-fir sites were 304 characteristically denser than the ponderosa pine sites, and therefore were already experiencing 305 more competition. The spruce-fir sites were at the stage of stand development where a second 306 phase stem exclusion process was becoming more pronounced. The first stage occurred earlier in 307 stand development, where overtopped or suppressed trees had died. Now, older trees were again 308 experiencing competition stress. As a result, successional changes were still occurring in the 309 spruce-fir stands whereas most of the ponderosa pine stands were older and showing less 310 successional changes. There was substantial mortality recorded in both pre and post-drought years, possibly a product of ongoing successional processes. As a whole, this ecozone did react to 312 the drought although in a more understated way when compared to ponderosa pine or the 313 piñon/juniper [1].

314
When the spruce-fir data was broken down into the two major habitat types, the Pien/Erex 315 habitat type showed some significant differences while the Abla/Vamy habitat type did not. The

322
The only significant differences in mortality for the individual species were the 0-10 cm in 323 Douglas-fir, 10-20 cm class in white fir, and 20-30 cm class in Engelmann spruce. There was 324 substantial pre-drought mortality, especially in the smaller size classes. This is most likely due to 325 the increased competition for resources in the stands and therefore increased stress on the trees.

326
The most interesting feature of the aspen data set is the large amount of mortality in both pre and

Conclusions
The significant differences in drought-induced mortality by the different size classes within the 362 different species were an important result of this study. As the region experiences a greater number 363 and intensity of drought events, major changes in the regional forest structure in the future most 364 likely will occur, with cascading impacts on wildlife, water, and the public.

369
Drought played a significant role in unthinned ponderosa pine mortality in Northern New Mexico.

370
This is no surprise as it has been estimated that snowpack in the Rockies can contain approximately showed no significance at all. This is an important implication because it shows that upon the onset 374 of a drought the overstory mortality can be reduced through mechanical thinning.

375
The drought increased all of the mortality percentages in the spruce-fir zone with the exception 376 of the two smallest size classes. The main difference between the ponderosa pine data and the 377 spruce-fir data is that the spruce-fir had substantial amounts of pre-drought mortality. The

378
Pien/Erex habitat type had significant increases in mortality suggesting that this habitat type is more 379 susceptible to drought.

390
We also thank anonymous reviewers for their recommendations to improve this manuscript.