1. Introduction
There have been major initiatives to restore longleaf pine (Pinus palustris Mill.) ecosystems, but little guidance has been given to landowners on strategies for adapting these ecosystems to predicted changes in temperature and precipitation. Across the Southern US, restoration activities have mostly focused on planting longleaf pine on marginal agricultural lands [1,2] and on cutover lands composed of other southern pines or hardwoods but with site characteristics that are more suitable for longleaf pine [3,4]. Other restoration activities include hardwood removal and the reintroduction of low-intensity surface fires in longleaf pine stands that are in degraded states, because fire suppression has led to the encroachment of mesic hardwoods and other pines that are less adapted to fire [5,6]. While tree planting and prescribed fire are often necessary components of longleaf pine restoration, the long-term success of these efforts should involve developing and evaluating structural and compositional targets for adapting longleaf pine ecosystems to climate change [7]. Findings from sites at which adaptive actions have been implemented are needed to inform broader use of adaptation approaches across longleaf pine ecosystems in the Southern US.
The Adaptive Silviculture for Climate Change (ASCC) project provides a framework for evaluating responses to climate change using a suite of adaptation approaches including resistance, resilience, and transition [7]. Our research focuses on the transition approach which aims to actively facilitate change to encourage adaptive responses. As part of this approach, forestry assisted migration (FAM) has been used to increase the proportion of tree species that are predicted to be better adapted to future climate conditions. The Jones Center at Ichauway, in Southwestern Georgia, includes one of fourteen ASCC core study sites across the USA and Canada. The Ichauway ASCC site is also one of five of the original ASCC sites, and it is the only ASCC site in the Southeastern US. During the planning phases of the Ichauway ASCC study site, the primary focus of managers and scientists was on drought impacts. Hence, the silvicultural prescription for the transition stands involved stand density reduction and species selection to reduce tree vulnerability to drought stress by decreasing competition for soil available water. Turkey oaks (Quercus laevis W.) were also planted as a part of an assisted population migration strategy [8]. This form of FAM is viewed as having the lowest amount of risk because it involves supplementing trees, typically from a more southerly latitude or lower elevation, to an existing population of the same species.
Tree planting that is part of FAM efforts can inform reforestation strategies that are mandated by the United States Congress. Such mandates include the prompt reforestation of national forestlands as outlined in The National Forest Management Act of 1976. The recent passage of the Repairing Existing Public Land by Adding Necessary Trees Act also increased annual funding for reforestation [9]. The success of FAM efforts could be based on tree vigor and survival with measurable ways to determine causes of poor tree health or mortality. For example, determining the overstory stand structure and species composition that promotes the early growth and survival of planted trees will help inform FAM efforts [10]. In the Southern US, there is a high degree of variability in site characteristics across longleaf pine ecosystems. Hence, studies of seedling performance across a variety of soil types would also provide guidance for choosing species that are part of FAM plantings. Lessons learned from case studies of FAM will be important for the broader use of FAM as a climate adaptation tool to minimize long-term risks to ecosystem vulnerability.
While FAM poses its own sets of challenges and risks [11,12], other components of the process of adapting stands to predicted climate conditions by intentionally facilitating change can present short-term risks. For example, reducing tree densities to low levels in transition stands at the Ichauway ASCC site unintentionally increased overstory tree vulnerability to wind disturbance. After the mortality and damage of trees due to Hurricane Michael, land managers and scientists at Ichauway had concerns about fuel discontinuity and the regeneration of longleaf pine. To address these concerns, evaluations of the percentage reduction in large overstory pines and stocking of natural longleaf pine regeneration are needed. While the assisted population migration of longleaf pine sourced from other seed zones or with genotypes that are particularly drought-tolerant may be an option for stands that lack a sufficient number of seed trees, longleaf pines were not planted in the transition stands at Ichauway. However, the FAM of longleaf pine will be important to consider because of the short-term risks associated with transition approaches and the ongoing restoration initiatives for planting on cutover and marginal agricultural lands.
The main objectives of this study were to determine (1) the survival of turkey oaks that were planted as part of an FAM strategy, and (2) the stocking of naturally regenerated longleaf pine in transition stands that were impacted by Hurricane Michael. The secondary objectives were to relate stand structure and species composition, as well as plant community and soil types, to regeneration outcomes. This research addresses several knowledge gaps about silvicultural approaches for adapting ecosystems to climate change that are based on experimental studies. To contribute to filling these knowledge gaps, we evaluated the resiliency of stands to disturbance, which was based on stand structure and regeneration outcomes. Our hope is that these findings will encourage land managers and researchers to consider implementing adaptation approaches across other areas of the Southern US as a way to maintain long-term ecosystem function, productivity, and services.
2. Materials and Methods
2.1. Study Area
The study site for this project was located at The Jones Center at Ichauway (31°19′ N, 80°20′ W), which is located in Southwestern Georgia, USA. Ichauway is within the southeastern Coastal Plain. The soils are primarily well-drained loamy sands over sandy loams. The climate of the region is characterized by long, hot summers and short, mild winters. This area of the coastal plain region features flat to gently rolling karst topography with low topographic relief. Most of 11,534 ha property was cut down in the early 1900s, resulting in the secondary growth of longleaf pine stands across much of the property.
In 2016, four blocks (A, B, C, and D) were delineated for the ASCC study at Ichauway. Each block contains five stands. Three adaptation approaches (resistance, resilience, and transition) and a control were assigned to stands within each block, and one stand was held in reserve. Before the ASCC treatments were initiated, all of the stands were composed of pines and hardwoods but were dominated by longleaf pines. Since the start of the study, prescribed fire has been used on a biennial basis during the dormant seasons in 2017, 2019, 2021, 2023, and 2025.
In 2017, eight 0.081 ha plots were established in each stand for measuring trees with diameter at breast height (DBH; stem diameter at 1.37 m) ≥ 12.7 cm. All trees were measured for DBH and height, and species were recorded. Inventories were conducted in 2017, after timber harvesting in 2018 but before Hurricane Michael, and every year thereafter until 2022. Hurricane Michael impacted Ichauway as a category 1 tropical cyclone in October of 2018.
Our study was limited to the transition stands. The objective of the transition approach was to reduce residual tree vulnerability to drought stress by reducing overstory tree densities and eliminating all mesic oaks and fleshy-fruited hardwoods. Tree cutting was carried out in January 2018, and target post-treatment basal areas for pine and residual oaks were 6.9 m2 ha−1 and 1.1 m2 ha−1, respectively. In 2022, xeric and upland oaks with DBH ≥ 12.7 cm included bluejack oak (Quercus incana Bartr.), turkey oak, sand post oak (Quercus margaretta Ashe), and southern red oak (Quercus falcata Mich.). Longleaf pine was the dominant pine, but a minor component of shortleaf pine (Pinus echinata Mill.) and slash pine (Pinus elliottii Engelm.) occurred in the transition stand of block D, and loblolly pine (Pinus taeda L.) was detected on some of the 0.081 ha plots in the transition stand of block B. Based on an examination of soil profiles by forest soil consultant John Holman (unpublished data including soil texture determined by feel, structure, consistency, and color of soil horizon), soil texture and horizon arrangements coincide with the Norfolk series for the transition stands in blocks A and B, the Troup series in block C, and the Troup and Wagram series in block D.
The turkey oaks were planted in the transition stands during March 2018. The one-year-old, bare-rooted seedlings were sourced from Superior Trees in Lee, Florida. The nursery is approximately 161 km southeast of Ichauway. Acorns used to propagate seedlings were likely collected in stands close to the nursery, but there are no records to verify this assumption. Six months prior to planting the turkey oaks at Ichauway, a circular area with a 2 m radius was treated with an herbicide mixture of Glypro (Glyphosate, Dow AgroSciences, Indianapolis, IN, USA) and Arsenal (Imazapyr, BASF Co., Ludwigshafen, Germany) at each planting location [13]. Each stand had a total of 150 planting locations, which were established using a 25 m2 systematic grid. At each planting location, six seedlings were planted in a circular pattern around a central seedling. The central seedling was marked with a metal tag on a curly stake that included the seedling plot number.
2.2. Data Collection
In May and June 2024, we conducted regeneration inventories in the four transition stands. In each stand, we attempted to locate the 150 clusters of turkey oak that were planted in 2018. Each cluster originally contained seven turkey oaks. When the planted turkey oaks or curly stake could be located (105, 72, 110, and 104 instances in blocks A, B, C, and D, respectively), we recorded the number of surviving turkey oaks. Potential reasons for why some clusters of turkey oaks could not be located were that logging trails and landings impacted planting locations and dense hardwood tickets established within the extent of other plantings. At the center of each cluster, we also measured the basal area of trees with diameter at breast height (DBH; stem diameter at 1.37 m) ≥ 12.7 cm using a wedge prism. We recorded total basal area, longleaf pine basal area, and hardwood basal area.
We also inventoried longleaf pine regeneration using a 40 m2 grid, which was different than the grid used to establish turkey oak planting locations. A separate grid system was used because herbicides were used to reduce understory competition in areas where turkey oaks were planted. Circular 0.001 ha plots were used to inventory regeneration. Longleaf pines < 2 m tall were tallied and two groundline diameters were recorded. Diameters were measured perpendicular to one another using calipers. We also recorded total height and height to the base of the apical bud, which were measured using a ruler. The presence of an overstory longleaf pine with a DBH ≥ 30 cm and within 30 m of the center of the regeneration plot, as well as the presence of wiregrass (Aristida stricta Michx.) within the regeneration plot, were also recorded.
2.3. Data Analyses
Generalized linear modeling was used to evaluate the influence of overstory hardwood basal area, overstory pine basal area, total overstory basal area, and block on planted turkey oak stocking. This modeling approach is commonly used for evaluating presence–absence data. Analysis of deviance tests were used to determine if explanatory variables should be retained in the binomial model [14]. Stocking was based on the presence of one live turkey oak per cluster. We used zero-inflated modeling to predict counts of turkey oak per cluster because the data included many zeros (i.e., locations with no surviving turkey oaks). A zero-inflated mixture model was deemed more appropriate than a hurdle model because detection of turkey oaks was not considered to be perfect [14]. This was due to the small stature of the planted turkey oaks, which were often hidden underneath tall grasses and scrubby oaks typical of two-year rough conditions (i.e., approximately 1.5 years since prescribed fire). The type of zero-inflated mixture model (zero-inflated Poisson (ZIP) model or zero-inflated negative binomial (ZINB) model) used in the analysis was based on overdispersion [14]. The pscl package [15] was used to fit these models. The package lmtest [16] was used to conduct a likelihood ratio test, which indicated that the ZIP model was the superior model. Next, we determined if explanatory variables (i.e., stand and basal area) could be dropped from the model. Reduced models were compared to one another and the full model. Models with the lowest Akaike information criterion (AIC) and p-values ≤ 0.05 were used as selection criteria to determine which explanatory variables to drop [14]. This process was repeated until all explanatory variables in the mixture model had p-values ≤ 0.05.
For longleaf pine regeneration with diameter at groundline > 1 cm, most of the non-zero counts were one. Hence, a simple binomial model was used to evaluate pine regeneration (i.e., counts were not modeled). Seedlings were placed into height classes with individuals < 25 cm tall being classified as in the “grass stage” and individuals ≥ 25 cm tall being classified as in the “bolting” or “rocket stage”. Separate models were created for each height class. Furthermore, seedlings < 25 cm tall were considered grass-stage individuals if they had an average groundline diameter > 1 cm. Following germination, longleaf pine juveniles are referred to as “pre-grass stage” for the first year or so of their growth and establishment [17].
3. Results
3.1. Stand Conditions
In October 2018, Hurricane Michael caused the partial mortality of and damage to overstory trees in the transition stands. The percent reduction in overstory longleaf pines with a DBH ≥ 28 cm on the 0.081 ha permanent plots was 6.3 ± 6.1% (grand mean ± standard deviation; stand means were used to calculate and report the grand mean and standard deviation). The greatest mean reduction in overstory longleaf pines was in the transition stand of block D (14.6%) (Figure 1). Before the hurricane, the post-treatment basal area and density of large longleaf pines were 6.0 ± 1.3 m2 ha−1 and 33 ± 9 trees ha−1, respectively. The percent reduction in longleaf pines with a DBH between 12.7 and 28 cm was 5.4 ± 13.6%. The greatest mean reduction in these smaller-sized longleaf pines was in the transition stand of block D (18.2%), and one transition stand had an increase in small longleaf pines due to ingrowth. Before the hurricane, the post-treatment density of small longleaf pines was 66 ± 24 trees ha−1. In 2024, the overstory longleaf pine and hardwood basal areas for trees with a DBH ≥ 12.7 cm associated with turkey oak regeneration plots were 7.6 ± 0.5 and 0.5 ± 0.2 m2 ha−1, respectively. Wiregrass occurred in 98% of the longleaf pine regeneration plots in the transition stands of block A and C. The groundcover of the transition stand in block D was dominated by species typically associated with disturbed soils (e.g., broomsedge bluestem (Andropogon virginicus L.)), and wiregrass only occurred in 8% of the regeneration plots. The transition stand of block B was composed of a mixture of native groundcover species and invasive plants (e.g., climbing fern (Lygodium japonicum (Thunb.) Sw.), and wiregrass occurred in 56% of the regeneration plots.
3.2. Turkey Oak Stocking
At least one live planted turkey oak was present in 74 ± 26% (grand mean ± standard deviation) of the planted clusters of turkey oaks that could be located six years after they were planted. The mean stocking for transition stands in blocks A, B, C, and D were 85, 35, 91, and 87%, respectively. The best model of turkey oak presence included the block and the overstory hardwood basal area as statistically significant explanatory variables (p < 0.05) (Table 1). Additionally, analyses of deviance suggested that these variables should be retained in the model. The overstory hardwood basal area and block explained 21.7% of the deviance in turkey oak presence. Deviance was calculated as the null deviance minus the residual deviance, which was then divided by the null deviance. The model indicated that the probability of turkey oak presence was greatest with low values for overstory hardwood basal area and in the transition stand of block C (Figure 2).
3.3. Turkey Oak Counts
In 2024, the number of live planted turkey oaks per cluster of seven turkey oaks that was planted in 2018 was 2.4 ± 1.2 (grand mean ± standard deviation). The mean counts per cluster for the transition stands in blocks A, B, C, and D were 3.0, 0.6, 3.1, and 2.9, respectively (Figure 3). The best ZIP model included the block and the overstory hardwood basal area (Table 2). The greatest probability for false zeros was obtained with high values for the overstory hardwood basal area and in block B (Figure 4). In other words, at these basal areas and in this block, we were likely to detect no planted turkey oaks in a cluster, but these zeros were false zeros. The predicted number of turkey oaks per cluster are displayed as red dots in Figure 3.
3.4. Longleaf Pine Stocking
At least one grass-stage longleaf pine was detected in 15 ± 12% (grand mean ± standard deviation) of the 0.004 ha regeneration plots (30 to 42 plots per stand). The mean stocking for transition stands in blocks A, B, C, and D was 11, 0, 26, and 23%, respectively. The optimal model of grass-stage longleaf pine presence included the block as a statistically significant explanatory variable. Additionally, analyses of deviance suggested that the block should be retained in the model, and overdispersion was minimal. The block explained 11.8% of the deviance in grass-stage longleaf pine presence. The model predicted the mean stocking for transition stands in blocks A, B, C, and D was 11, 0, 26, and 23%, respectively.
At least one rocket-stage longleaf pine was detected in 9 ± 5% (grand mean ± standard deviation) of the regeneration plots. The mean stocking for transition stands in blocks A, B, C, and D was 13, 3, 7, and 13%, respectively. Analyses of deviance suggested that the block should be dropped from the model of rocket-stage longleaf pine presence.
4. Discussion
4.1. Transition Approaches
Among the different adaptation strategies, the transition approach seeks to facilitate and promote compositional and structural conditions that are anticipated to occur with changes in climate and disturbance regimes. The silvicultural actions implemented in the transition stands at Ichauway, which included reducing overstory densities to extremely low levels and assisted population expansion planting, inherently introduce upfront risks in the short term. In contrast, a resistance approach seeks to maintain current conditions [7]. This may be beneficial in the short term but could potentially lead to risks and increased costs of maintaining desired conditions in the long term because of changes in climate and disturbance regimes. While the transition stands of this study were somewhat susceptible to the hurricane, such risks are expected to diminish over time as residual trees become more windfirm and additional trees are recruited to larger size classes. Although eight 0.081 ha plots were established in each stand for measuring trees with a DBH ≥ 12.7 cm, we also acknowledge that the spatial heterogeneity of overstory trees within the transition stands could have led to an underestimation or overestimation of the mean reduction rates in longleaf pines after Hurricane Micheal. On a relative basis (i.e., the percent reduction), our results also suggest that large pines with a DBH ≥ 28 cm were just as susceptible to the hurricane as smaller longleaf pines. The relatively small reduction in longleaf pines across size classes indicates that the stands were resilient to a category 1 tropical cyclone.
4.2. Turkey Oak Regeneration
Zero-inflated modeling was especially appropriate for our count data, yielding more accurate predictions of planted turkey oak presence and revealing trends that were not obvious at the start of the study. Previous research conducted in the transition stands suggested that the physiological performance of the planted turkey oaks was influenced by competitive pressure from the overstory, but only in the more mesic stands [13]. In our study, seedling survival was negatively correlated with overstory hardwood basal area, which provides additional evidence that the overstory may influence planted turkey oak persistence in these stands. The transition stands in blocks A, C, and D had a greater mean stocking of planted turkey oaks than that of the transition stand of block B. The stand in block B is located adjacent to the Flint River corridor and likely has soils with greater clay contents than the other transition stands. The soils of the stand were best described as coinciding with the Norfolk series, which typically has a Bt horizon that is classified as having a sandy clay loam texture. Furthermore, the presence of non-native invasive species in this stand is likely because of the stand’s proximity to the Flint River. In our zero-inflated model, the dense understory of native and invasive species likely explains the greater probability of false zeros (i.e., the chance of not detecting small-statured turkey oak seedlings when, in fact, they are there) in this stand compared to the other transition stands at Ichauway. The relatively low amount of variation that was explained in the turkey oak stocking by the overstory hardwood basal area and block (21.7%) suggests that other factors such as fire behavior, understory competition, and soil properties may have influenced turkey oak survival at fine spatial scales within the stands.
For the purpose of the ASCC study, turkey oak stocking across blocks A, C, and D was deemed to be sufficient. We do not expect every planted individual to survive, but that enough individuals persist as a component of the understory layer, with some individuals eventually reaching the midstory and reproductive maturity. Other ASCC sites have also implemented FAM plantings using a suite of species and genotypes [10,18,19]. For three ASCC sites in Minnesota and New England, Palik et al. [10] found no consistent trends in seedling survival between different types of FAM (assisted population expansion, range expansion, and species migration) among sites. Five years after planting at the Minnesota Red Pine ASCC site, approximate survival rates for assisted population expansion plantings ranged from 50%–70% survival among different combinations of oak species and seed sources [19]. At our study site, the planted turkey oaks in transition stands of blocks A, C, and D will help to gradually facilitate structural and compositional changes. In the transition stand of block B, we may consider planting other species that are more appropriate for the soil types of the block. Our study suggests the importance of considering potential variability in site conditions at established and future ASCC sites. Researchers can be limited when selecting treatment locations to maintain uniform site conditions, so careful selection of the species used in FAM plantings is important. These findings also demonstrate the need for timely monitoring to assess and potentially adapt silvicultural actions in order to meet desired future conditions.
4.3. Longleaf Pine Regeneration
Limited information exists in the scientific literature about what is considered adequate stocking for longleaf pine regeneration in uneven-aged stands. For recommended seedling densities after overstory removals, Croker and Boyer [20] and Brockway et al. [21] have suggested that about 14,800 seedlings (greater than one year old but less than one meter tall) per hectare would result in about 1235 well-distributed saplings (equal or greater than 1 m tall) per hectare. In mature longleaf pine flatwood ecosystems, Gagnon et al. [22] described a mean regeneration stocking of 16% across canopy gaps as “low stocking”, a value similar to the mean values obtained in our study. Both Croker and Boyer [20] and Brockway et al. [21] note that seedling stocking can and should be dependent on specific management objectives. While our mean stocking values may seem low, for the purpose of maintaining uneven-aged stand conditions, we do not expect or require 100% stocking across every part of the stand. It is also important to note that the transition approach is a relatively new climate-adaptive strategy, so there is not yet information on appropriate stocking levels. Because the purpose of the transition approach is to maintain low stand densities to increase water yields, we consider the longleaf pine regeneration across blocks A, C, and D to be satisfactory. Also, the low stand densities provide growing space for additional regeneration over time. The low stocking levels detected in the transition stand of block B could be due to the same factors that likely reduced turkey oak survival: a greater soil clay content compared to other transition stands, as well as understory competition, especially from non-native invasive species. In the transition stand of block B, managers could consider supplementing additional natural regeneration by timing a late-summer prescribed burn to coincide with a “bumper cone crop” or heavy masting event.
We also intended to examine the influence of seed trees and understory community type on longleaf pine regeneration. Croker and Boyer [20] described the presence of a seed source as an important factor in natural longleaf pine regeneration. We recorded the presence of at least one large overstory longleaf pine tree within 20 m of the plot center. However, we detected such trees near every plot across all four blocks. In southwestern ponderosa pine (Pinus ponderosa C. Lawson var. scopulorum Engelm.) forests, Puhlick et al. [23] found that the herbaceous understory community type had an influence on pine regeneration. We recorded the understory community type, but we found that it was highly correlated with the stand location (i.e., the block random effect in our models). Hence, this variable was also excluded from our modeling efforts. Future research could consider other potentially influential characteristics like soil moisture or competition from other tree species in the understory. The continual monitoring of regeneration and stand structure and species composition over time will be important.
5. Conclusions
The use of a transition approach is novel within the historic range of longleaf pine, and the ASCC study at Ichauway provides a means to evaluate treatment outcomes. Post-treatment stand conditions potentially increase the stand’s susceptibility to major disturbance events (e.g., hurricanes), which could alter ecosystem functionality. Our findings demonstrate the ability of transition stands to accommodate a large-scale disturbance event (i.e., Hurricane Michael) and maintain ecosystem functionality, the desired stand structure, and species composition. The novel FAM plantings of climate-adapted species also come with concerns about perceptions, risks, and maladaptation [10]. The relative success of FAM plantings of turkey oaks at our ASCC site may alleviate some concerns of maladaptation. Our study also highlights the need for the timely monitoring and adaptation of treatment actions as researchers and practitioners learn more about climate-focused FAM implementation. Despite the upfront risks, transition approaches present a viable option as climate adaptive strategies, with their risks expected to diminish over time.
A.S.H.: conceptualization; data curation; formal analysis; investigation; methodology; supervision; validation; visualization; writing—original draft; writing—review and editing. J.J.P.: conceptualization; data curation; formal analysis; funding acquisition; investigation; methodology; project administration; resources; software; supervision; validation; visualization; writing—original draft; writing—review and editing. All authors have read and agreed to the published version of the manuscript.
The data presented in this study are available on request from the corresponding author. The data are not publicly available due to privacy as other reports are currently being drafted with the data.
We would like to thank Allyson Horn, Hallie Turner, Ben Campbell, and Scott Taylor for their assistance with the fieldwork. We would like to thank Krishna Poudel (Mississippi State University) for reviewing an early draft of this manuscript. We would like to thank Linda Nagel (Utah State University), Courtney Peterson and Maria Vicini (Colorado State University), and Maria Janowiak (USDA Forest Service) for their discussions about the Ichauway ASCC site.
The authors declare that this study received funding from Ichauway Inc. The funder was not involved in the study design, collection, analysis, interpretation of data, the writing of this article or the decision to submit it for publication.
Footnotes
Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content.
Figure 1 Residual stand conditions after hurricane in block D (a). Planted turkey oak as part of a forest assisted migration strategy (b). Wiregrass-dominated groundcover in block A (c), and a mixture of native and invasive species in block B (d). Photos taken in 2022, courtesy of J.J.P.
Figure 2 Predicted probabilities (solid lines) and 95% confidence bands (dashed lines) of planted turkey oak presence six years after planting turkey oaks in the transition stands of the ASCC study at Ichauway.
Figure 3 Observed counts of live planted turkey oaks per cluster in 2024, as well as the Pearson residuals from the zero-inflated model. The boxes define the interquartile range (IQR, 25–75 percent quartile), and boxes are proportional to the number of observations per block. The vertical line below the boxes extends to a horizontal line that is the 25th percentile minus 1.5 times the IQR, and the vertical line above the boxes extends to a horizontal line that is the 75th percentile minus 1.5 times the IQR. In each box, the horizontal line is the median, the black dot is the observed mean, and the red dot is the zero-inflated model’s predicted mean.
Figure 4 Fitted curves for the logistic regression model of turkey oak presence showing the probability of obtaining a false zero.
Parameter estimates and model fit statistics for the model of turkey oak presence that contained block (categorical; A, B, C, D) and overstory hardwood basal area (Hwd BA; m2 ha−1).
| Model 1 | ai (SE) | |||
|---|---|---|---|---|
| Block A | Block B | Block C | Block D | |
| ai −0.26681 (Hwd BA) | 1.95110 (0.29585) | −0.52034 (0.37972) | 2.41012 (0.43690) | 2.02955 (0.40193) |
| Slope SE | Null Deviance | Residual Deviance | ||
| 0.10176 | 414.51 | 324.67 | ||
1 ln (probability of turkey oak presence). SE, standard error.
Coefficients for the count model (Poisson with log link) and the zero-inflated model (binomial with logit link) that contained block (categorical; A, B, C, D) and overstory hardwood basal area (Hwd BA; m2 ha−1).
| Model | Hwd BA | Block A | Block B | Block C | Block D |
|---|---|---|---|---|---|
| Count | NA | 1.23301 | −0.94068 | −0.04770 | −0.07151 |
| Zero-inflated | −0.30516 | −2.2533 | 2.25205 | −0.72192 | −0.14811 |
NA, not applicable; Hwd BA was not included in the count model.
1. Barbarika, A. Farm Service Agency. The Conservation Reserve Program: A 35-Year History; USDA: Washington, DC, USA, 2021; 21.
2. Puhlick, J.J.; Markewitz, D.; Taylor, R.S. Soil change after planting longleaf pine on agricultural lands: Revisiting a chronosequence study after 23 years. Soil Sci. Soc. Am. J.; 2025; 89, e70062. [DOI: https://dx.doi.org/10.1002/saj2.70062]
3. Kush, J.S.; Meldahl, R.S.; McMahon, C.K.; Boyer, W.D. Longleaf pine: A sustainable approach for increasing terrestrial carbon in the southern United States. Environ. Manage.; 2004; 33, pp. S139-S147. [DOI: https://dx.doi.org/10.1007/s00267-003-9124-3]
4. Puhlick, J.J.; O’Halloran, T.L.; Starr, G.; Abney, R.B.; Knapp, L.P.; McCleery, R.A.; Klepzig, K.D.; Brantley, S.T.; McIntyre, R.K.; Song, B. Opportunities for research on carbon management in longleaf pine ecosystems. Forests; 2023; 14, 874. [DOI: https://dx.doi.org/10.3390/f14050874]
5. Condon, B.; Putz, F.E. Countering the broadleaf invasion: Financial and carbon consequences of removing hardwoods during longleaf pine savanna restoration. Restor. Ecol.; 2007; 15, pp. 296-303. [DOI: https://dx.doi.org/10.1111/j.1526-100X.2007.00212.x]
6. Kirkman, L.K.; Giencke, L.M.; Taylor, R.S.; Boring, L.R.; Staudhammer, C.L.; Mitchell, R.J. Productivity and species richness in longleaf pine woodlands: Resource-disturbance influences across an edaphic gradient. Ecology; 2016; 97, pp. 2259-2271. [DOI: https://dx.doi.org/10.1002/ecy.1456] [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/27859094]
7. Nagel, L.M.; Palik, B.J.; Battaglia, M.A.; D’Amato, A.W.; Guldin, J.M.; Swanston, C.W.; Janowiak, M.K.; Powers, M.P.; Joyce, L.A.; Millar, C.I.
8. Bower, A.D.; Frerker, K.L.; Pike, C.C.; Labonte, N.R.; Palik, B.J.; Royo, A.A.; Anderson, S.M.; Ferreira, A.R.; Brandt, L.A. A practical framework for applied forestry assisted migration. Front. For. Glob. Change; 2024; 7, 1454329. [DOI: https://dx.doi.org/10.3389/ffgc.2024.1454329]
9. Balloffet, N.; Dumroese, R.K. The national reforestation strategy and the REPLANT Act: Growing and nurturing resilient forests. Foundational Concepts in Silviculture with Emphasis on Reforestation and Early Stand Improvement—2022 National Silviculture Workshop Proceedings; Jain, T.B.; Schuler, T.M. USDA Forest Service, Rocky Mountain Research Station: Fort Collins, CO, USA, 2022; 3.
10. Palik, B.J.; Clark, P.W.; D’Amato, A.W.; Swanston, C.; Nagel, L. Operationalizing forest-assisted migration in the context of climate change adaptation: Examples from the eastern USA. Ecosphere; 2022; 13, e4260. [DOI: https://dx.doi.org/10.1002/ecs2.4260]
11. Pedlar, J.H.; McKenney, D.W.; Aubin, I.; Beardmore, T.; Beaulieu, J.; Iverson, L.; O’Neill, G.A.; Winder, R.S.; Ste-Marie, C. Placing forestry in the assisted migration debate. Bioscience; 2012; 62, pp. 835-842. [DOI: https://dx.doi.org/10.1525/bio.2012.62.9.10]
12. Williams, M.I.; Dumroese, R.K. Preparing for climate change: Forestry and assisted migration. J. For.; 2013; 111, pp. 287-297. [DOI: https://dx.doi.org/10.5849/jof.13-016]
13. Reynolds, S.A. Turkey oak (Quercus laevis W.) seedlings experience primarily competitive pressures, not facilitation, from the overstory in a longleaf pine (Pinus palustris M.) ecosystem. Master’s Thesis; Auburn University: Auburn, AL, USA, 2020.
14. Zuur, A.F.; Ieno, E.N.; Walker, N.J.; Saveliev, A.A.; Smith, G.M. Mixed Effects Models and Extensions in Ecology with R; Springer: New York, NY, USA, 2009.
15. Zeileis, A.; Kleiber, C.; Jackman, S. Regression models for count data in R. J. Stat. Softw.; 2008; 27, pp. 1-25. [DOI: https://dx.doi.org/10.18637/jss.v027.i08]
16. Zeileis, A.; Hothorn, T. Diagnostic checking in regression relationships. R. News; 2002; 2, pp. 7-10.
17. Keeley, J.E. Ecology and evolution of pine life histories. Ann. For. Sci.; 2012; 69, pp. 445-453. [DOI: https://dx.doi.org/10.1007/s13595-012-0201-8]
18. Crotteau, J.S.; Sutherland, E.K.; Jain, T.B.; Wright, D.K.; Jenkins, M.M.; Keyes, C.R.; Nagel, L.M. Initiating climate adaptation in a western larch forest. For. Sci.; 2019; 65, pp. 528-536. [DOI: https://dx.doi.org/10.1093/forsci/fxz024]
19. Palik, B.J.; Kastendick, D.N.; Kragthorpe, J. Comparing performance of assisted migration seed sources of two oak species in a Minnesota red pine woodland. New For.; 2024; 55, pp. 1875-1885. [DOI: https://dx.doi.org/10.1007/s11056-024-10064-8]
20. Croker, T.C., Jr.; Boyer, W.D. Regenerating Longleaf Pine Naturally; USDA Forest Service, Southern Forest Experiment Station: New Orleans, LA, USA, 1975; 26.
21. Brockway, D.G.; Outcalt, K.W.; Boyer, W.D. Longleaf pine regeneration ecology and methods. Ecology, Silviculture, and Restoration; Shibu, J.; Jokela, E.J.; Miller, D.L. Springer: New York, NY, USA, 2006; pp. 95-133.
22. Gagnon, J.L.; Jokela, E.J.; Moser, W.K.; Huber, D.A. Characteristics of gaps and natural regeneration in mature longleaf pine flatwoods ecosystems. For. Ecol. Manage.; 2004; 187, pp. 373-380. [DOI: https://dx.doi.org/10.1016/j.foreco.2003.07.002]
23. Puhlick, J.J.; Laughlin, D.C.; Moore, M.M. Factors influencing ponderosa pine regeneration in the southwestern USA. For. Ecol. Manage.; 2012; 264, pp. 10-19. [DOI: https://dx.doi.org/10.1016/j.foreco.2011.10.002]
© 2025 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https://creativecommons.org/licenses/by/4.0/). Notwithstanding the ProQuest Terms and Conditions, you may use this content in accordance with the terms of the License.
Details
1 The Jones Center at Ichauway, Newton, GA 39870, USA; [email protected], College of Forest Resources, Mississippi State University, Starkville, MS 39762, USA
2 The Jones Center at Ichauway, Newton, GA 39870, USA; [email protected]