Abstract

Through modification of structural characteristics, ecological processes such as fire can affect microhabitat parameters, which in turn can influence community composition dynamics. The prevalence of high-severity forest fires is increasing in the southern and western United States, creating the necessity to better understand effects of high-severity fire, and subsequent postfire management actions, on forest ecosystems. In this study we used a recent high-severity wildfire in the Lost Pines ecoregion of Texas to assess effects of the wildfire and postfire clearcutting on six microclimate parameters: air temperature, absolute humidity, mean wind speed, maximum wind speed, soil temperature, and soil moisture. We also assessed differences between burned areas and burned and subsequently clearcut areas for short-term survivorship of loblolly pine Pinus taeda seedling trees. We found that during the summer months approximately 2 y after the wildfire, mean and maximum wind speed differed between unburned and burned areas, as well as burned and burned and subsequently clearcut areas. Our results indicated air temperature, absolute humidity, soil temperature, and soil moisture did not differ between unburned and burned areas, or burned and burned and subsequently clearcut areas, during the study period. We found that short-term survivorship of loblolly pine seedling trees was influenced primarily by soil type, but was also lower in clearcut habitat compared with habitat containing dead standing trees. Ultimately, however, the outcome of the reforestation initiative will likely depend primarily on whether or not the trees can survive drought conditions in the future, and this study indicates there is flexibility in postfire management options prior to reseeding. Further, concerns about negative wildfire effects on microclimate parameters important to the endangered Houston toad Bufo (Anaxyrus) houstonensis were not supported in this study.

Introduction

Species occurrence and abundance at both site and landscape scales is influenced by abiotic environmental parameters (e.g., temperature and humidity), either directly through species preferences and tolerance limits, or indirectly through habitat effects (e.g., soil characteristics) and competition dynamics (Neuenschwander and Wright 1984; Baltz et al. 1987; Adolph 1990; Chase and Leibold 2003). In forest ecosystems, ecological processes such as fire can have strong effects on microhabitat parameters, representing one way in which ecological processes can affect community composition dynamics (Walker et al. 2003; North et al. 2005; Ma et al. 2010). Microhabitat impacts occur through modifications of structural characteristics that influence these parameters, such as canopy cover and depth of ground litter (Ripley and Archibold 1999; Rocca 2009).

The climatic pattern toward warmer and drier conditions, coupled with longstanding broad-scale fire suppression, have resulted in an increase in high-severity wildfires (i.e., wildfires that kill or top-kill the majority of live vegetation and consume the majority of dead organic matter) in the southern and western United States (Davis 2001; Crotteau et al. 2013), particularly in forest ecosystems (Miller et al. 2009; Liu et al. 2013), with this pattern projected to continue into the next century (Moritz et al. 2012). Thus, increasing our understanding of impacts of high-severity fire, and subsequent postwildfire management actions, on forest ecosystems is currently of high interest (Bisson et al. 2003; Beschta et al. 2004; Cairney and Bastias 2007). Postfire management actions vary from allowing natural recovery and regeneration to high levels of manipulation, such as salvage logging and active revegetation (Driscoll et al. 2010; Powers et al. 2013). Management actions are chosen based on goals associated with the unique ecological, geographic, and social–political characteristics of a burned area, perceived action results, and available funding for restoration (Agee 1996; Driscoll et al. 2010).

In September and October 2011, a high-severity wildfire occurred in the Lost Pines ecoregion of Texas, burning approximately 39% of the 34,400-ha ecoregion. With respect to property loss, the fire was regarded as the most destructive wildfire occurring in state history for an urban–wildland interface environment (Ridenour et al. 2012). Postfire management decisions were driven by the complex social and ecological context of the Lost Pines. The majority of burned land in the ecoregion was privately owned (76.3%), and property impacts were substantial (e.g., 1,696 structures burned; Lost Pines Recovery Team 2011; Ridenour et al. 2012). Thus, the majority of the postfire management decisions were focused on mitigation of hazards to humans. Ecologically, the Lost Pines is a disjunct loblolly pine Pinus taeda forest that represents the westernmost extent of the species distribution in the United States, and which is a genetically distinct population (Al-Rabah'ah and Williams 2004). In addition, the ecoregion is the last remaining stronghold for the federally endangered Houston toad Bufo (Anaxyrus) houstonensis, and thus wildlife management actions are focused largely on this species (Brown et al. 2011; Duarte et al. 2011).

Several large tracts (i.e., >1000-ha) consisting of primarily undeveloped land of high conservation value currently exist within the Lost Pines ecoregion. These tracts are critical for preserving the ecological integrity of the ecoregion given ongoing substantial habitat fragmentation of parcels surrounding the tracts. Two properties in particular, the Griffith League Ranch (GLR; 1,948-ha) and Bastrop State Park (BSP; 2,398-ha), are considered essential for long-term persistence of the Houston toad in the ecoregion (Hatfield et al. 2004). These two properties are currently taking two different postfire management strategies. Because natural loblolly pine regeneration was low and patchy across most of the burn zone, assisted pine restoration through seedling tree planting is currently occurring on both properties (Lost Pines Recovery Team 2012). However, the GLR strategy includes removal of all dead standing trees prior to planting; whereas, with the exception of trees adjacent to roads, dead standing trees are being left on BSP. It is currently unclear which of these strategies is most appropriate for restoration of the Lost Pines ecoregion.

The purpose of this study was to document short-term impacts of the wildfire, and subsequent management actions, on microclimate parameters that likely effect survivorship of both loblolly pines and Houston toads, and Houston toad breeding activity. In addition, we investigated the impact of the two management actions on short-term survivorship of loblolly pine seedling trees. The information obtained from this study will be useful for guiding future forest-recovery actions in the Lost Pines, as well as in comparable forest ecosystems where managers are dealing with similar restoration decisions.

Study Site

This study was conducted on the GLR and BSP in the Lost Pines ecoregion in Bastrop County, Texas, USA (Figure 1). The Lost Pines is a remnant patch of pine-dominated forest that is thought to have been isolated from the East Texas Piney Woods ecoregion between 10,000 and 14,000 y ago (Bryant 1977), with the pines of the area beginning genetic differentiation from those populations further east by 30,000 y ago (Al-Rabah'ah and Williams 2004). Both the GLR and BSP are primarily forested properties, with an overstory dominated by loblolly pine, post oak Quercus stellata, and eastern red cedar Juniperus virginiana. The properties are primarily underlain by deep sandy soils of the Patilo–Demona–Silstid Association, with shallow gravelly soils also present on the southern end of BSP (Baker 1979).

Figure 1.

Aerial image of the Griffith League Ranch (GLR; northern property outline) and Bastrop State Park (BSP; southern property outline), Bastrop County, Texas, and their location with respect to wildfires that occurred in the Lost Pines ecoregion in September and October 2011, which burned approximately 13,500 ha. Overlain on the image are the pond and seedling tree sampling areas used to study the impacts of the wildfire and postfire management actions on microclimate parameters, loblolly pine Pinus taeda seedling tree survivorship, and endangered Houston toad Bufo (Anaxyrus) houstonensis predicted calling activity in spring and summer 2013.

Figure 1.

Aerial image of the Griffith League Ranch (GLR; northern property outline) and Bastrop State Park (BSP; southern property outline), Bastrop County, Texas, and their location with respect to wildfires that occurred in the Lost Pines ecoregion in September and October 2011, which burned approximately 13,500 ha. Overlain on the image are the pond and seedling tree sampling areas used to study the impacts of the wildfire and postfire management actions on microclimate parameters, loblolly pine Pinus taeda seedling tree survivorship, and endangered Houston toad Bufo (Anaxyrus) houstonensis predicted calling activity in spring and summer 2013.

On 4 September 2011 a high-severity wildfire (the Bastrop County Complex Fire) began from multiple initial fire outbreaks across the Lost Pines. The fire was unstoppable because of wind gusts in excess of 58 km/h resulting from the passage of tropical storm Lee, coupled with extreme drought conditions in central Texas (based on the Palmer Drought Severity Index; Lost Pines Recovery Team 2011). After 18 d the fire was 95% contained, with the total burn area encompassing 13,406 ha. A fire break was installed on the GLR during the burn, restricting the fire on the property to 987 ha (50% of the GLR). On 4 October 2011, the fire breached the break and burned an additional 81 ha. In comparison, the September fire burned nearly the entirety of BSP (approx. 95% of the property). The estimated overstory tree loss was 78% across the entirety of the Bastrop County Complex Fire area (Lost Pines Recovery Team 2011). Based on 13 randomly placed 20 m × 50 m vegetation plots on the GLR within the wildfire zone, the mean overstory tree loss was 81.9%, with a corresponding 30.2% decrease in overstory canopy cover, approximately 1 y after the wildfire (Brown et al. 2014).

Methods

We conducted this study between 11 April 2013 and 23 June 2013, approximately 2 y after the high-severity wildfire. We assessed wildfire and postfire management impacts to six microclimate parameters: air temperature, absolute humidity (estimated from temperature and relative humidity [Brown et al. 2013]), mean wind speed, maximum wind speed, soil temperature (2–3 cm below the soil surface), and soil moisture (2–3 cm below the soil surface). We obtained parameter estimates using Davis Vantage Pro2 portable weather stations with leaf and soil moisture–temperature stations (Davis Instruments, Hayward, CA). We used two portable weather stations, which allowed us to collect data at control and treatment (i.e., unburned, burned, and clearcut) locations simultaneously. Although the weather stations were precalibrated by the manufacturer, prior to the study we placed them adjacent to one-another and assessed data collected over several days to ensure estimates were similar and no directional bias was apparent. For each pair of observations, we allowed stations to collect data for 13–190 h (i.e., 26–380 data points/station; mean  =  30.2 h). We logged data every 30 min, with each data point representing the mean value from the 30-min sampling interval, with the exception of maximum wind speed, which represented the maximum value recorded over the 30-min sampling interval.

This study consisted of three components. First, we investigated microclimate differences between unburned and burned habitat at 14 known Houston toad breeding ponds (Brown et al. 2013; 7 in burned areas and 7 in unburned areas), and their surrounding uplands. Secondly, we investigated microclimate differences between burned habitat and burned habitat that was subsequently clearcut (hereafter called clearcut; Figure 2). Lastly, we investigated differences in short-term survivorship of loblolly pine seedling trees between burned and clearcut habitat, as well as between two soil types (deep sandy soil and shallow gravelly soil).

Figure 2.

Portable weather stations used to assess impacts of a high-severity wildfire that occurred in summer 2011 on microclimate parameters in spring and summer 2013 on the Griffith League Ranch (GLR), Bastrop County, Texas, USA. The left panel (A) shows an example of burned forest, and the right panel (B) shows an example of burned forest that was subsequently clearcut.

Figure 2.

Portable weather stations used to assess impacts of a high-severity wildfire that occurred in summer 2011 on microclimate parameters in spring and summer 2013 on the Griffith League Ranch (GLR), Bastrop County, Texas, USA. The left panel (A) shows an example of burned forest, and the right panel (B) shows an example of burned forest that was subsequently clearcut.

To investigate microclimate differences between unburned and burned habitat at Houston toad breeding ponds and their surrounding uplands, we sampled two random points around each pond (i.e., within 3 m of the pond edge), with corresponding sampling points placed 50 m into the surrounding upland. We randomized sampling points using a geographic information system (GIS; ArcGIS 10.1; ESRI, Redlands, CA). Specifically, using the GIS we created two random points within each pond polygon. We then walked along the edge of each pond with a portable GPS unit (GPSMAP 60 CSx; Garmin, Olathe, KS), and located the point along the pond edge that minimized the straight-line distance to each random point specified by the GIS.

We investigated microclimate differences between burned and clearcut habitat shortly after the clearcutting operation began because of concerns about potential negative impacts of clearcutting on the Houston toad through microclimate impacts (e.g., increased potential for desiccation due to decreased soil moisture). Thus, data for the burned habitat versus clearcut habitat component of this study was limited to an 11-ha clearcut patch, and we assumed data collected in this patch was representative of patches that would subsequently be clearcut. For this comparison, we selected 10 random points in the clearcut zone, with the restriction that points be ≥50 m from one another and ≥20 m from adjacent burned forest. We paired the 10 clearcut sampling points with points in adjacent burned forest, with stations placed 50 m from, and perpendicular to, the forest edge closest to the clearcut sampling points. In addition, we estimated overstory canopy cover at each weather station sampling point using a spherical densiometer.

To assess short-term (i.e., 3–5 mo postplanting) loblolly pine seedling survivorship, we sampled burned and clearcut zones in two soil types, one consisting of deep sandy soil and one consisting of shallow gravelly soil. For the deep sandy soil habitat, we sampled 10 areas in burned (BSP) and clearcut (GLR) habitat, using 6 m × 30 m plots placed at randomly selected points and azimuths. For the shallow gravelly soil habitat (BSP), we sampled 10 areas in burned and clearcut habitat on BSP. Because both the clearcut and seedling planting occurred in a strip along a park road, we placed plots perpendicular to, and approximately 20 m from, the clearcut edge, at randomly selected points along the road. Thus, this data set contained 4 categories with 10 samples each: burned–deep sandy soil, burned–shallow gravelly soil, clearcut–deep sandy soil, and clearcut–shallow gravelly soil.

Planted seedling trees were approximately 1 y old (J. M. Rooni, Texas A&M Forest Service, personal communication), and typically >13 cm tall (personal observation; Figure 3). On BSP flags were placed adjacent to planted trees during planting events, and during this study we found a living or dead tree next to all but two flags, indicating disappearance of trees after planting was minimal. Although flags were not placed adjacent to planted trees on the GLR, we likely detected most of the seedling trees because of their size and because the recent clearcutting resulted in minimal vegetation cover at the time of planting and, subsequently, sampling.

Figure 3.

Loblolly pine Pinus taeda seedling trees used for a large-scale seedling tree-planting initiative after a high-severity fire in the Lost Pines ecoregion, Bastrop County, Texas, in summer 2011 that killed approximately 78% of overstory trees across 13,487 ha. Image used with permission from photographer Jim Rooni.

Figure 3.

Loblolly pine Pinus taeda seedling trees used for a large-scale seedling tree-planting initiative after a high-severity fire in the Lost Pines ecoregion, Bastrop County, Texas, in summer 2011 that killed approximately 78% of overstory trees across 13,487 ha. Image used with permission from photographer Jim Rooni.

We used two-stage analyses to compare microclimates in unburned, burned, and clearcut habitats. Specifically, we compared unburned vs. burned habitat adjacent to ponds, unburned vs. burned habitat in the surrounding uplands, and burned vs. clearcut habitat. In the first stage, we used a multivariate analysis approach to test for an overall difference (typically referred to as a community-level difference), and inferred responses of individual variables using ordination diagrams. This is a useful data exploration and reduction tool when the number of response variables is large (McCune and Grace 2002). In the second stage, we used a univariate analysis approach to investigate differences for individual variables that appeared to differ between habitat types based on the ordination diagrams.

For the multivariate analyses, we used redundancy analysis (RDA), which is an extension of principal components analysis (PCA) to include explanatory variables. Specifically, for RDA each response variable is regressed on each explanatory variable, and then a PCA is performed on the matrix of fitted values (McCune and Grace 2002). We chose RDA (which assumes response variables are related linearly to predictors) over canonical correspondence analysis (which assumes response variables are related unimodally to predictors) because our gradient lengths were short (<4) and our predictors were categorical (Lepš and Šmilauer 2003). We tested for a treatment effect using a Monte Carlo permutation test, with data randomized within, but not among, treatment-control pairs. We centered and standardized the response data because variables were measured on different scales (i.e., response variables had a zero average and unit variance). We additionally log10-transformed the response data, so that percentages rather than absolute changes were analyzed (note this transformation applies only to the RDA analyses). We assessed individual variable responses using species-environment biplots. We performed these analyses using the program CANOCO (version 4.5).

For the univariate analyses we used linear mixed-effects models, specifying habitat category as a fixed effect, and replicate (i.e., pond, surrounding upland, or burned and/or clearcut sampling point) as a random effect (Pinheiro and Bates 2000). For each analysis we assessed assumptions of normality and homoscedasticity using graphical diagnostic plots (Zuur et al. 2009), and transformed variables when applicable using the arcsinh (i.e., inverse hyperbolic sine) transformation (Fowler et al. 1998). We performed these analyses using the nlme package (version 3.1-113) in the program R (version 3.0.2).

To assess seedling pine tree survivorship differences between habitat types (burned vs. clearcut) and soil types (sandy vs. gravelly), we used generalized linear mixed-effects models with a binomial distribution, the Gauss–Hermite approximation to the log-likelihood (nAGQ  =  9), and plot specified as a random effect (Bates et al. 2013). To determine the most parsimonious model, we used a likelihood ratio test and compared four competing models: the full model (Treatment + Soil + Treatment × Soil), the additive effects model (Treatment + Soil), and models with only soil type or treatment included (Zuur et al. 2009). For the most parsimonious model, we assessed significance of coefficients using the Wald Z-statistic (Bolker et al. 2009). We performed these analyses using the glmer function for the lme4 package (version 1.0-5) in the program R (version 3.0.2). For all analyses we considered differences to be significant at α  =  0.1.

Results

Mean percent canopy cover was 38.1% lower in burned habitat (mean  =  35.6% ± 31.6%) compared with unburned habitat (mean  =  57.5% ± 34.3%) for ponds, and 51.6% lower in burned habitat (mean  =  26.2% ± 25.4%) compared with unburned habitat (mean  =  54.1% ± 35.5%) for surrounding uplands. Clearcutting on the 11-ha forested area on the GLR reduced mean percent canopy cover from 40.3% ± 19.7% to 0% (Table S1).

The RDA analyses indicated overall differences between burned and unburned areas surrounding ponds (P  =  0.001), and in adjacent uplands (P  =  0.001), as well as between clearcut and burned areas (P  =  0.001; Table S2). Further, biplots for each analysis indicated potential differences in mean and maximum wind speed for all comparisons, soil temperature for the pond and upland comparisons, and soil moisture for the upland and clearcut comparisons (Figure 4).

Figure 4.

Results from Redundancy Analyses (RDA) used to assess the impacts of a high-severity wildfire that occurred in summer 2011 on microclimate parameters in spring and summer 2013 on the Griffith League Ranch (GLR), Bastrop County, Texas, USA. The top panel (A) compared unburned and burned habitat around endangered Houston toad Bufo (Anaxyrus) houstonensis breeding ponds. The middle panel (B) compared unburned and burned habitat in adjacent surrounding uplands. The bottom panel (C) compared burned and burned and subsequently clearcut habitat. Direction and strength of associations between microclimate parameters and habitat types are indicated by the direction (e.g., a microclimate parameter arrow pointing approx. 180° from a habitat category arrow would indicate a strong negative association), and length (longer arrows indicate stronger associations), of the arrows. Additional univariate analyses indicated only mean and maximum wind speed differed between compared habitat types (P-values in the figure are from additional univariate analyses).

Figure 4.

Results from Redundancy Analyses (RDA) used to assess the impacts of a high-severity wildfire that occurred in summer 2011 on microclimate parameters in spring and summer 2013 on the Griffith League Ranch (GLR), Bastrop County, Texas, USA. The top panel (A) compared unburned and burned habitat around endangered Houston toad Bufo (Anaxyrus) houstonensis breeding ponds. The middle panel (B) compared unburned and burned habitat in adjacent surrounding uplands. The bottom panel (C) compared burned and burned and subsequently clearcut habitat. Direction and strength of associations between microclimate parameters and habitat types are indicated by the direction (e.g., a microclimate parameter arrow pointing approx. 180° from a habitat category arrow would indicate a strong negative association), and length (longer arrows indicate stronger associations), of the arrows. Additional univariate analyses indicated only mean and maximum wind speed differed between compared habitat types (P-values in the figure are from additional univariate analyses).

For the pond comparison, the univariate analyses indicated mean wind speed (F1,12  =  2.76, P  =  0.123), maximum wind speed (F1,12  =  2.13, P  =  0.118), and soil temperature did not differ significantly (F1,12  =  0.37, P  =  0.553). For the upland comparison, the univariate analyses indicated mean (F1,12  =  6.83, P  =  0.023) and maximum (F1,12  =  6.47, P  =  0.026) wind speed was higher in the burned treatment, but soil temperature (F1,12  =  2.61, P  =  0.146) and moisture (F1,12  =  0.37, P  =  0.554) did not differ significantly. For the clearcut comparison, the univariate analyses indicated mean (F1,18  =  3.09, P  =  0.096) and maximum (F1,18  =  4.10, P  =  0.058) wind speed was higher in the clearcut treatment, but soil moisture did not differ significantly (F1,18  =  0.59, P  =  0.452).

We detected 1,096 seedling pine trees in the 40 plots (Table S3). Mean survivorship per plot was 0.59 (SD  =  0.21) and 0.68 (SD  =  0.18) for clearcut and burned habitat in the deep sandy soil type, respectively. Mean survivorship per plot was 0.95 (SD  =  0.07) and 0.99 (SD  =  0.03) for clearcut and burned habitat in the gravelly soil type, respectively. The likelihood-ratio test indicated the most parsimonious model included habitat type and soil type as predictors, with no habitat type × soil type interaction (χ21  =  48.96, P < 0.001). The coefficient estimates and Wald Z-statistics indicated soil type was the most important seedling pine tree survivorship predictor (βSandy soil  =  −3.31, Z  =  −8.22, P < 0.001), but habitat type was also important (βClearcut  =  −0.69, Z  =  −1.94, P  =  0.053).

Discussion

Despite decreases in overstory canopy cover, the results of this study indicated that, with the exception of wind speed, microclimate parameters were similar between unburned and burned habitat, and burned and clearcut habitat. This is a positive result with respect to concerns about negative wildfire impacts on microclimate parameters important for Houston toad survivorship (Brown et al. 2014). Further, with respect to Houston toad calling activity, we used a previously published statistical model to estimate calling probability during the sampling period (11 April–10 May 2013) at each of the sampled ponds (Brown et al. 2013), and a paired randomization test with 10,000 iterations to determine whether calling probability differed between unburned and burned ponds (n  =  7 pairs), which indicated no difference in calling probability between the two habitat types (P  =  0.261; Table S4). However, we note that higher wind speeds in burned and clearcut habitat could negatively affect Houston toads through increased potential for desiccation while individuals are on the surface (Hillman et al. 2009).

For loblolly pine seedling trees, soil type had the strongest influence on survivorship. We believe this is likely due to the shallow gravelly soil type holding more rain water near the surface, whereas water probably drained below root complexes in the deep sandy soil type (Baker 1979). During federal recovery operations, private landowners excavated many stumps from the local loblolly pine mature stands in the deep sandy soil type. We noted during operations that these stumps provided a shallow fine-root system near the ground surface, columnar structural roots, and then a secondary horizontal spread of fine roots, with the deeper roots all at approximately the same depth (1.0 m ± 0.16 m). We speculate that this relates not only to structural capability, but also water availability at this depth in the deep sandy soil type. If this hypothesis is correct, replanting success for loblolly pine trees in the deep sandy soil type will depend heavily on rainfall levels in the coming months and years. Bastrop County was under severe drought conditions in spring and summer 2013 (based on the Palmer Drought Severity index), but returned to nondrought conditions in autumn 2013, and was only abnormally dry (least severity level) in early 2014 (Available: http://www.drought.gov, accessed 29 November 2013).

Although soil type was clearly a more important factor affecting loblolly pine seedling tree survivorship, our analyses indicated short-term survivorship was lower in clearcut compared with burned habitat (i.e., in clearcut areas, seedling tree survivorship was approximately 9% lower in sandy soil and 4% lower in gravelly soil). However, the results of this study did not provide evidence that microclimate components important for loblolly pine survivorship differed between burned and clearcut habitat in the sandy soil type. Thus, it is currently unclear why clearcutting negatively affected survivorship. One possibility is that soil disturbance from tractors and skid steers during clearcutting altered soil properties that we did not measure, such as nutrient availability and compaction, and that could impact seedling tree survival (Simard et al. 2001; Simcock et al. 2006; Thiffault et al. 2008). Ultimately, however, the outcome of the reforestation initiative will likely depend primarily on whether or not the trees can survive drought conditions in the future, and this study indicates there is flexibility in postfire management options prior to reseeding.

Our long-term concern for both western (i.e., more drought-adapted) loblolly pine and the Houston toad is their capability to adapt to climatic changes in the southern United States, which are resulting in warmer and drier conditions (Liu et al. 2013). The U.S. Department of Agriculture Forest Service's Climate Change Tree Atlas indicates the Lost Pines ecoregion may be unsuitable habitat for loblolly pine by 2100 under both high- and low-emissions scenarios (Prasad et al. 2007–ongoing). This issue is already becoming apparent, because an estimated 5.7% of overstory loblolly pine trees (i.e., trees that had survived approx. 20–60+ y) died during exceptional drought conditions in 2009 (based on the Palmer Drought Severity index; Brown et al. 2013). In addition, terrestrial mesocosm experiments indicate desiccation is a major source of mortality for Houston toads during drought conditions (M. Jones, Texas State University, personal communication). This emphasizes the need to provide data enabling management decisions to achieve the goal of restoring this forest system as efficiently and cost-effectively as possible.

Supplemental Material

Please note: The Journal of Fish and Wildlife Management is not responsible for the content or functionality of any supplemental material. Queries should be directed to the corresponding author.

Table S1. Overstory canopy-cover estimates at portable weather-station points used to determine whether microclimate parameters differed between unburned habitat, habitat that was burned during a high-severity wildfire in September 2011, and habitat that was burned and subsequently clearcut, on the Griffith League Ranch (GLR) in the Lost Pines ecoregion, Bastrop County, Texas, USA, in summer 2013. The Comparison column delineates the component of this study each point was used for, and the Habitat column delineates the habitat category for each point.

Found at DOI: 10.3996/072013-JFWM-051.S1 (10.4 KB XLSX)

Table S2. Microclimate parameter estimates from portable weather stations used to determine whether microclimate parameters differed between unburned habitat, habitat that was burned during a high-severity wildfire in September 2011, and habitat that was burned and subsequently clearcut, on the Griffith League Ranch (GLR) in the Lost Pines ecoregion, Bastrop County, Texas, USA, in summer 2013.

Found at DOI: 10.3996/072013-JFWM-051.S2 (336 KB XLSX)

Table S3. Seedling loblolly pine Pinus taeda tree survivorship data used to determine whether survivorship was associated with treatment (burned vs. clearcut) and soil type (sandy vs. gravelly) on the Griffith League Ranch (GLR) and Bastrop State Park (BSP) in the Lost Pines ecoregion, Bastrop County, Texas, USA, in summer 2013.

Found at DOI: 10.3996/072013-JFWM-051.S3 (12 KB XLSX)

Table S4. Data used to determine whether predicted calling activity of the endangered Houston toad Bufo (Anaxyrus) houstonensis on the Griffith League Ranch (GLR) in the Lost Pines ecoregion, Bastrop County, Texas, USA, differed between areas burned in a 2011 high-severity wildfire and unburned areas, in summer 2013.

Found at DOI: 10.3996/072013-JFWM-051.S4 (63.6 KB XLSX)

Acknowledgments

We thank G. Creacy and the Texas Parks and Wildlife Department for allowing us to use Bastrop State Park for this study, and for giving us necessary Geographic Information System files to complete this study. The Boy Scouts of America provided access to the Griffith League Ranch and we are grateful for their support. We thank A. Villamizar and A. Parandhaman for assistance with field data collection. Three anonymous reviewers and the Subject Editor provided suggestions that greatly improved the quality of this manuscript. This study was supported by Texas State University-San Marcos through a doctoral research stipend, and the Texas Parks and Wildlife Department and U.S. Fish and Wildlife Service through a traditional Section 6 grant.

Any use of trade, product or firm names is for descriptive purposes only and does not imply endorsement by the U.S. Government.

References

Adolph
SC
.
1990
.
Influence of behavioral thermoregulation on microhabitat use by two Sceloporus lizards
.
Ecology
71
:
315
327
.
Agee
JK
.
1996
.
Achieving conservation biology objectives with fire in the Pacific Northwest
.
Weed Technology
10
:
417
421
.
Al-Rabah'ah
MA
,
Williams
CG
.
2004
.
An ancient bottleneck in the Lost Pines of central Texas
.
Molecular Ecology
13
:
1075
1084
.
Baker
FE
.
1979
.
Soil survey of Bastrop County, Texas
.
Washington, D.C.
:
U.S. Soil Conservation Service and Texas Agricultural Experiment Station
.
Baltz
DM
,
Vondracek
B
,
Brown
LR
,
Moyle
PB
.
1987
.
Influence of temperature on microhabitat choice by fishes in a California stream
.
Transactions of the American Fisheries Society
116
:
12
20
.
Bates
D
,
Maechler
M
,
Bolker
B
,
Walker
S
.
2013
.
Package ‘lme4’
. .
Beschta
RL
,
Rhodes
JJ
,
Kauffman
JB
,
Gresswell
RE
,
Minshall
GW
,
Karr
JR
,
Perry
DA
,
Hauer
FR
,
Frissell
CA
.
2004
.
Postfire management on forested public lands of the western United States
.
Conservation Biology
18
:
957
967
.
Bisson
PA
,
Rieman
BE
,
Luce
C
,
Hessburg
PF
,
Lee
DC
,
Kershner
JL
,
Reeves
GH
,
Gresswell
RE
.
2003
.
Fire and aquatic ecosystems of the western USA: current knowledge and key questions
.
Forest Ecology and Management
178
:
213
229
.
Bolker
BM
,
Brooks
ME
,
Clark
CJ
,
Geange
SW
,
Poulsen
JR
,
Stevens
MHH
,
White
JS
.
2009
.
Generalized linear mixed models: a practical guide for ecology and evolution
.
Trends in Ecology and Evolution
24
:
127
135
.
Brown
DJ
,
Baccus
JT
,
Means
DB
,
Forstner
MRJ
.
2011
.
Potential positive effects of fire on juvenile amphibians in a southern USA pine forest
.
Journal of Fish and Wildlife Management
2
:
135
145
.
Brown
DJ
,
Nowlin
WH
,
Ozel
E
,
Mali
I
,
Episcopo
D
,
Jones
MC
,
Forstner
MRJ
.
2014
.
Comparison of short term low, moderate, and high severity fire impacts to aquatic and terrestrial ecosystem components of a southern USA mixed pine/hardwood forest
.
Forest Ecology and Management
312
:
179
192
.
Brown
DJ
,
Swannack
TM
,
Forstner
MRJ
.
2013
.
Predictive models for calling and movement activity of the endangered Houston toad
.
American Midland Naturalist
169
:
303
321
.
Bryant
VM
Jr.
1977
.
A 16,000 year pollen record of vegetational change in central Texas
.
Palynology
1
:
143
156
.
Cairney
JWG
,
Bastias
BA
.
2007
.
Influences of fire on forest soil fungal communities
.
Canadian Journal of Forest Research
37
:
207
215
.
Chase
JM
,
Leibold
MA
.
2003
.
Ecological niches: linking classical and contemporary approaches
.
Illinois
:
University of Chicago Press
.
Crotteau
JS
,
Varner
JM
III,
Ritchie
MW
.
2013
.
Post-fire regeneration across a fire severity gradient in the southern Cascades
.
Forest Ecology and Management
287
:
103
112
.
Davis
C
.
2001
.
The west in flames: the intergovernmental politics of wildfire suppression and prevention
.
Publius
31
:
97
110
.
Driscoll
DA
,
Lindenmayer
DB
,
Bennett
AF
,
Bode
M
,
Bradstock
RA
,
Cary
GJ
,
Clarke
MF
,
Dexter
N
,
Fensham
R
,
Friend
G
,
Gill
M
,
James
S
,
Kay
G
,
Keith
DA
,
MacGregor
C
,
Possingham
HP
,
Russel-Smith
J
,
Salt
D
,
Watson
JEM
,
Williams
D
,
York
A
.
2010
.
Resolving conflicts in fire management using decision theory: asset-protection versus biodiversity conservation
.
Conservation Letters
3
:
215
223
.
Duarte
A
,
Brown
DJ
,
Forstner
MRJ
.
2011
.
Estimating abundance of the endangered Houston toad on a primary recovery site
.
Journal of Fish and Wildlife Management
2
:
207
215
.
Fowler
J
,
Cohen
L
,
Jarvis
P
.
1998
.
Practical statistics for field biology. 2nd edition
.
New York
:
John Wiley & Sons
.
Hatfield
JS
,
Price
AH
,
Diamond
DD
,
True
CD
.
2004
.
Houston toad (Bufo houstonensis) in Bastrop County, Texas: need for protecting multiple subpopulations
. Pages
292
298
in
Akcakaya
HR
,
Burgman
MA
,
Kindvall
O
,
Wood
CC
,
Sjogren-Gulve
P
,
Hatfield
JS
,
McCarthy
MA
,
editors
.
Species conservation and management: case studies
.
Oxford, England, United Kingdom
:
Oxford University Press
.
Hillman
SS
,
Withers
PC
,
Drewes
RC
,
Hillyard
SD
.
2009
.
Ecological and environmental physiology of amphibians
.
Oxford, England, United Kingdom
:
Oxford University Press
.
Lepš
J
,
Šmilauer
P
.
2003
.
Multivariate analysis of ecological data using CANOCO
.
Cambridge, England, United Kingdom
:
Cambridge University Press
.
Liu
Y
,
Goodrick
SL
,
Stanturf
JA
.
2013
.
Future U.S. wildfire potential trends projected using a dynamically downscaled climate change scenario
.
Forest Ecology and Management
294
:
120
135
.
Lost Pines Recovery Team
.
2011
.
Bastrop County complex fire, Lost Pines region, resources assessment and response report
. .
Lost Pines Recovery Team
.
2012
.
Bastrop Lost Pines habitat recovery project master plan
. .
Ma
SY
,
Concilio
A
,
Oakley
B
,
North
M
,
Chen
J
.
2010
.
Spatial variability in microclimate in a mixed-conifer forest before and after thinning and burning treatments
.
Forest Ecology and Management
259
:
904
915
.
McCune
B
,
Grace
JB
.
2002
.
Analysis of ecological communities
.
Gleneden Beach, Oregon
:
MjM Software Design
.
Miller
JD
,
Safford
HD
,
Crimmins
M
,
Thode
AE
.
2009
.
Quantitative evidence for increasing forest fire severity in the Sierra Nevada and Southern Cascade mountains, California and Nevada, USA
.
Ecosystems
12
:
16
32
.
Moritz
MA
,
Parisien
M-A
,
Batllori
E
,
Krawchuk
MA
,
Van Dorn
J
,
Ganz
DJ
,
Hayhoe
K
.
2012
.
Climate change and disruptions to global fire activity
.
Ecosphere 3:article 49
.
Neuenschwander
LF
,
Wright
HA
.
1984
.
Edaphic and microclimate factors affecting tobosagrass regrowth after fire
.
Journal of Range Management
37
:
116
122
.
North
M
,
Hurteau
M
,
Fiegener
R
,
Barbour
M
.
2005
.
Influence of fire and El Niño on tree recruitment varies by species in Sierran mixed conifer
.
Forest Science
51
:
187
197
.
Pinheiro
JC
,
Bates
DM
.
2000
.
Mixed-effects models in S and S-PLUS
.
New York
:
Springer
.
Powers
EM
,
Marshall
JD
,
Zhang
JW
,
Wei
L
.
2013
.
Post-fire management regimes affect carbon sequestration and storage in a Sierra Nevada mixed conifer forest
.
Forest Ecology and Management
291
:
268
277
.
Prasad
AM
,
Iverson
LR
,
Matthews
S
,
Peters
M
.
2007–ongoing
.
A climate change atlas for 134 forest tree species of the eastern United States [database]
.
U.S. Department of Agriculture Forest Service. Available: http://www.nrs.fs.fed.us/atlas/tree (November 2013)
.
Ridenour
K
,
Rissel
S
,
Powell
W
,
Gray
R
,
Fisher
M
,
Sommerfeld
J
.
2012
.
Bastrop complex wildfire case study
.
Texas Forest Service and Bastrop County Office of Emergency Management. Available: http://www.co.bastrop.tx.us/bcdisaster/index.php/wildfire-case-study (July 2013)
.
Ripley
EA
,
Archibold
OW
.
1999
.
Effects of burning on prairie aspen grove microclimate
.
Agriculture Ecosystems & Environment
72
:
227
237
.
Rocca
ME
.
2009
.
Fine-scale patchiness in fuel load can influence initial post-fire understory composition in a mixed conifer forest, Sequoia National Park, California
.
Natural Areas Journal
29
:
126
132
.
Simard
DG
,
Fyles
JW
,
Paré
D
,
Nguyen
T
.
2001
.
Impacts of clearcut harvesting and wildfire on soil nutrient status in the Quebec boreal forest
.
Canadian Journal of Soil Science
81
:
229
237
.
Simcock
RC
,
Parfitt
RL
,
Skinner
MF
,
Dando
J
,
Graham
JD
.
2006
.
The effects of soil compaction and fertilizer application on the establishment and growth of Pinus radiata. 2006
.
Canadian Journal of Forest Research
36
:
1077
1086
.
Thiffault
E
,
Hannam
KD
,
Quideau
SA
,
Paré
D
,
Bélanger
N
,
Oh
S-W
,
Munson
AD
.
2008
.
Chemical composition of forest floor and consequences for nutrient availability after wildfire and harvesting in the boreal forest
.
Plant Soil
308
:
37
53
.
Walker
LR
,
Lodge
DJ
,
Guzmán-Grajales
SM
,
Fetcher
N
.
2003
.
Species-specific seedling responses to hurricane disturbance in a Puerto Rican rain forest
.
Biotropica
35
:
472
485
.
Zuur
AF
,
Ieno
EN
,
Walker
NJ
,
Saveliev
AA
,
Smith
GM
.
2009
.
Mixed effects models and extensions in ecology with R
.
New York
:
Springer Science+Business Media
.

Author notes

Brown DJ, Mali I, Forstner MRJ. 2014. Wildfire and postfire restoration action effects on microclimate and seedling pine tree survivorship. Journal of Fish and Wildlife Management 5(1):174-182; e1944-687X. doi: 10.3996/072013-JFWM-051

The findings and conclusions in this article are those of the author(s) and do not necessarily represent the views of the U.S. Fish and Wildlife Service.