Free

The Intersection between Climate Adaptation, Mitigation, and Natural Resources: An Empirical Analysis of Forest Management

and

Abstract

Forest landowners can adapt to climate change and carbon pricing by altering the types of forests that are replanted or regenerated. By inducing land-use changes within forestry, climate adaptation and mitigation policy can alter the flow of nonmarket forest ecosystem services. The purpose of this paper is to quantify the effect of climate change and carbon pricing on adaptation behavior of private forest owners. We develop an empirical framework with application to the US Pacific coast. An estimated discrete-choice econometric model is used as the empirical basis for a simulation of land-use changes to the composition of a landscape’s forest stock. Results indicate that climate change induces landowners to adapt away from their current dominant species choice of Douglas-fir to species more suitable for the future climate, notably hardwoods and ponderosa pine. A carbon price policy accelerates adaptation away from current forest types, potentially creating an externality at the local level.

Climate adaptation, mitigation, and natural resources interact in numerous ways that generate social costs from climate change. In the case of privately owned forest resources, climate change impacts on forest growth and disturbance can induce landowners to adapt through management adjustments on the intensive margin (e.g., altering the timing and intensity of harvests) and through extensive margin changes in tree species planting (Guo and Costello 2013). Forest management changes—especially on the extensive margin—result in land-use changes to the composition of a landscape’s forest stock, thereby altering the flow of a landscape’s ecosystem services. For example, changes in forest management can (i) alter the rate of carbon sequestration on a landscape, (ii) produce different water quality outcomes (Fulton and West 2002), and (iii) alter biodiversity through habitat changes that affect wildlife who are habitat specialists (Wilcove et al. 1998). The potential link between climate adaptation, forest composition, and ecosystem services suggests that privately optimal adaptation in forestry will interact with social nonmarket values from forests, and this interaction will generate trade-offs that affect the social cost of climate change. How climate adaptation, forest composition, and ecosystem services interact is not well understood in the current literature on climate change and natural resources.

A complicating factor in studying the effects of climate adaptation on forest composition and ecosystem services is the question of how climate mitigation policy may induce additional adaptations by landowners. A carbon price policy aimed at mitigating climate change creates carbon rents that vary across alternative forest types depending on their sequestration rates (Ekholm 2016). A carbon price can either accelerate or push back on extensive margin adjustments and, therefore, accelerate or push back on adaptive land-use changes within forestry. A carbon rent implicitly rewards the planting of tree species that generate the highest flow of one ecosystem service (carbon sequestration) and potentially at the expense of other tree species that may generate higher valued flows of other nonmarket ecosystem services (e.g., wildlife habitat, water quality, etc.). Pricing only one ecosystem service may, therefore, generate important trade-offs with other ecosystem services that flow from forests. While previous research has identified a similar trade-off between encouraging carbon sequestration and preserving biodiversity across broad land uses (e.g., Nelson et al. 2008; Lawler et al. 2014), little is known about how forest land-use decisions are affected by climate change and how the composition of forests would be influenced by the interaction between climate adaptation and carbon pricing. The possibility that carbon pricing could induce adaptive land-use and habitat changes within forestry means that carbon pricing could induce unintended negative externalities on nonmarket ecosystem service flows within forests. The social optimality of carbon pricing is far from clear when there are interactions between nonmarket services and feedbacks from the nonmarket consequences of policy interventions (Carbone and Smith 2013). A necessary step in understanding whether the social optimality of carbon pricing may be affected by interactions with nonmarket natural resources is to examine the magnitude of joint adaptation to climate change and carbon pricing on physical stocks of natural resources.

This paper develops an empirical framework to model adaptation decisions to climate change and a carbon price by private forest landowners, with application to the Pacific states of the United States—California, Oregon, and Washington. We use plot-level data to empirically estimate a discrete-choice econometric model of management choices as a function of timber prices, yields, site productivity, and measures of downscaled climate that correspond to the plot. Our key source for identifying climate adaptation in forestry is to exploit spatial variation in climatic variables and replanting choices across recently harvested timber plots, controlling for a set of fine-scale information on key drivers of rents from replanting or regenerating specific forest types. The econometric estimates define a set of plot-level probabilities of key forest management decisions that are used to simulate landscape change that results from a series of changes in multiple climate variables and a hypothetical carbon-pricing scheme. Beginning with the current forest stock on the landscape, the simulation uses the econometric estimates to adjust plot-level management probabilities to exogenous changes in climate. The simulation generates endogenous changes in the forest stock, including the timing and intensity of harvest, natural disturbance, and the composition of different forest types in repeated 10-year intervals until the year 2100. Results show that along the US Pacific coast, landowners gradually shift out of their current dominant species choice of Douglas-fir to species more suitable for the future climate, notably hardwoods and ponderosa pine. Results also show that a carbon price policy would further accelerate adaptation away from existing Douglas-fir stocks. Since many local wildlife species of conservation concern are specialized to Douglas-fir rather than hardwood or ponderosa forests, a carbon price aimed at internalizing a global externality may generate localized externalities by increasing the speed of land-use and habitat changes arising from extensive margin adaptations.

Recent literature that emphasizes the benefits of using empirically derived human-climate linkages has focused on quantifying the economic damage from climate change on the value of agricultural land (Schlenker et al. 2006), labor markets (Graff Zivin and Neidell 2014), and electricity demand (Auffhammer et al. 2017). Notably absent from the current empirical literature are estimates of climate’s impact on the market value of forestland and nonmarket changes in biodiversity and ecosystem services (Carleton and Hsiang 2016). Moreover, despite the potential consequences of adaptation behaviors on ecosystem services, there has been little empirical analysis of indirect damage from climate change that operates through adaptation behavior, and especially how mitigation policy interacts with adaptation. By showing the possibility that adaptation behavior and mitigation policy reinforce local land-use change effects from climate change, our analysis contributes to the policy discussion about the relative roles of adaptation and mitigation (Fankhauser 2017).

We also contribute to a small but growing list of recent literature that extends econometric land-use models (e.g., Lubowski et al. 2006) to examine climate-driven land use change and its effects on ecosystem services (Fezzi et al. 2015; Bateman et al. 2016). Fezzi et al. (2015) examine the problem of deterioration of river water quality in the United Kingdom due to land-use change as a result of climate adaptation in the farming sector, and they consider a potential policy response to the adaptation-induced environmental problem. In contrast, our study focuses on how a policy aimed at mitigating climate change damages may actually contribute to local land-use changes that can deteriorate ecosystem service provision. Further, by modeling forest management in response to climate, we contribute to the natural science literature that studies climate change impacts on forest resources (e.g., Coops and Waring 2011; Hanewinkel et al. 2012; Iverson and McKenzie 2013; Prasad et al. 2013; Rehfeldt et al. 2014; Mathys et al. 2017).

Finally, our approach of empirically estimating adaptation behavior shares similarities with the literature on agriculture and climate change that has studied adaptation implicitly through the effects of climate change on land prices (Mendelsohn et al. 1994; Schlenker et al. 2006; Deschênes and Greenstone 2007; Severen et al. 2018), and explicitly modeled the choice of crops to plant as a function of climate (Seo and Mendelsohn 2008). There are few econometric studies of climate adaptation in forests, although Guo and Costello (2013) and Hannah et al. (2011) use numerical dynamic programming techniques to examine the value of adaptation and how climate change can affect forest structure in California through privately optimal adaptation. Our approach is inspired by the theoretical framework advanced by Guo and Costello (2013), and we contribute to this earlier work by econometrically estimating adaptation behavior.

1. Overview of the Framework

1.1. Conceptual Research Design

Our modeling goal is to develop empirical evidence that allows us to analyze changes in the forest composition of a landscape under climate change and carbon-pricing scenarios, relative to a baseline. Landscape change in forest composition results from forest-management decisions of many individual landowners, where each individual landowner chooses management to maximize the value of his forest given exogenous timber prices, biophysical conditions of the plot (e.g., soil), the state of the forest stand, climate, and unobserved heterogeneous preferences.

Figure 1 illustrates the conceptual framework of the econometric-based simulation for a given forest landowner of a timber plot. The landowner treats timber prices, soils, and climate as exogenous attributes that influence his management choices. The landowner of a stand planted with species s_j in age a first chooses whether to harvest his land as a clear-cut (remove all timber volume), a partial-cut (remove a partial amount of volume), or no-cut (no harvest). We use an econometric model to parameterize a probabilistic function of the landowner’s choice of the three harvest possibilities as a function of exogenous attributes. If the landowner harvests his land as a clear-cut, he then chooses the tree species with which to replant or regenerate his land; if the landowner harvests his land as a partial-cut, he chooses which tree species remains on the land to naturally regenerate other trees. For the replanting/regeneration choices, we use an econometric model to parameterize a probabilistic function of the landowner’s choice of which tree species to replant/regenerate as a function of exogenous attributes.¹ If the landowner chooses not to harvest his land and let it grow, then he faces the possibility that his stand will be naturally disturbed by fire, disease, or insect damage. We also use an econometric model to parameterize a probabilistic binary function of natural disturbance as a function of exogenous attributes. If the stand is undisturbed, we use empirically calibrated timber yield functions to determine how much timber volume grows to the next time period. The econometric parameters are derived by simultaneously estimating harvest choices, replanting choices, and natural disturbance outcomes in a nested framework.

After the harvest, replanting, and natural disturbance outcomes are determined in a given period, we then update the forest attributes of the stand (species type, age, growth, volume) and move to the next time period when the harvest decision is revisited. We repeatedly revisit the harvest decision in 10-year increments, t, $t + 10$ , $t + 20$ , and so on for each plot. Scaling up the management choices of all landowners within a landscape generates the composition of the landscape across different types of forests. By repeating the forest management decisions many times in a Monte Carlo style (app. B) that accounts for the probabilistic nature of the econometric model, we generate distributions of landscape change under alternative scenarios (Lewis and Plantinga 2007).

We use our landscape simulation to estimate the effects of climate change and carbon pricing on the forest composition of the landscape. In each 10-year increment, a baseline scenario updates timber prices using global price projections without climate change while holding all other climate variables fixed. Our climate change scenario updates climate variables according to climate change forecasts, timber prices using global price projections under climate change, and the net primary productivity of forests as estimated by natural scientists under climate change. Our carbon-pricing scenario introduces a carbon payment based on the sequestration path arising from the state of the stand (forest type, site class, age) and offered to the landowner as a rental payment for their carbon sequestration.

Key to our research design is the fact that the decision rules (probabilities) that drive forest management choices are estimated using a discrete-choice econometric framework under the assumption that landowners reveal their optimal management choice through their observed management choices. Since econometric partial effects do not directly indicate landscape changes resulting from forest management, the landscape simulation is used to translate a series of spatially heterogeneous changes in multiple downscaled climate variables and a hypothetical carbon-pricing scheme into landscape changes that are consistent with our econometric evidence. Estimation uses spatial variation in climate and recent forest management decisions based on the USDA Forest Service Forest Inventory and Analysis (FIA) as the basis for estimation. Figure 2 illustrates a basic empirical link between climate on the US Pacific coast and the existing forest types on the landscape. The data in figure 2 come from linking observable locations of existing forest plots in the FIA data to long-run climate averages at that plot. However, because forests are stocks, the existing forest represents a culmination of a set of adaptations that have been occurring for decades—for example, a 40-year-old Douglas-fir stand resulted from a landowner’s replanting choice 40 years ago. In contrast, our econometric model is not based on the existing forest stock but, rather, on observed management choices during the period 2001–14.

**Figure 2.**
Relationship between current distribution of tree species and climate. Values after each species name represent average temperature and total precipitation over the sampled plots during the growing season in the current climate.
View Large Image Download PowerPoint

1.2. Relationship to Literature on Economics of Forest Supply

Our modeling framework builds off of, and is differentiated from, at least two major strands of economics literature that models timber production and supply. First, there is a literature that uses market simulation models to solve for the dynamic path of equilibrium price and quantity of timber that maximizes the present value of the total surplus from producing and consuming timber products (Lyon and Sedjo 1983). Market models have been extensively used to examine long-run timber supply (Adams et al. 1996), the impacts of climate change on timber markets (Sohngen and Mendelsohn 1998; Lee and Lyon 2004; Sohngen and Tian 2016), and the effects of carbon prices on markets (Im et al. 2007). Market simulation model results find that climate change will be beneficial to timber markets (Mendelsohn et al. 2016). The strength of market simulation models is their internal structural consistency in capturing the dynamic equilibrium effects of impacts from supply shocks. Critiques of simulation models include the assertion that they “rely too heavily on assumptions rather than empirical facts” (Massetti and Mendelsohn 2018, 327) and that by imposing a single objective function on landowners, they fail to link aggregate supply to heterogeneous individual harvest behavior (Polyakov et al. 2010) and do not account for unobservable heterogeneity in landowner preferences (Stavins 1999).

Second, there is a literature that uses discrete-choice econometric methods to estimate the effects of economic and plot-level factors on the timber harvest choice at the plot level. This literature commonly uses plot-level observations of harvest choice as the dependent variable (e.g., Dennis 1990; Provencher 1997; Prestemon and Wear 2000; Polyakov et al. 2010). Plot-level econometric studies assume that prices and yields are exogenous, and timber supply can be constructed by aggregating the estimated plot-level choices at a range of simulated exogenous timber prices (Prestemon and Wear 2000; Polyakov et al. 2010). While the econometric approach overcomes the above critiques of simulation models by constructing empirical evidence based on the revealed behavior of landowners, a weakness of the econometric approach is the inability to characterize market equilibrium in a dynamically consistent fashion.

Our econometric framework contributes to the above literature by estimating the effects of climate on discrete plot-level timber management choices and by adapting the econometric-based simulation approach used in land-use modeling to examine the effects of climate change and carbon pricing on forest composition. Our approach to nesting harvest and replanting choices as a function of climate extends prior forestry studies of supply. By basing our model on revealed management choices across space, our approach is a cross-sectional approach to studying climate adaptation (Massetti and Mendelsohn 2018). Previous literature that examines the interaction between climate and forestry is predominantly based on dynamic market simulation models that assume rather than estimate adaptation (Sohngen and Mendelsohn 1998; Lee and Lyon 2004; Sohngen and Tian 2016).

2. Theoretical Basis for Forest Management

This section and the next section extend Guo and Costello’s (2013) theoretical work by showing how a nested logit discrete choice econometric framework can be used to estimate the basic relationships between climate and management behaviors. Our focus is on developing a positive analysis of landowner behavior that is driven by observed management decisions.

Consider a forest landowner who has just chosen a harvest method h and is now choosing which forest type to replant post-harvest. If a landowner chooses the clear-cut harvest method, then his discrete-choice problem is to choose the forest type s_j to replant. Alternatively, a landowner who recently conducted a partial-cut harvest now owns a stand with potentially mixed ages, described by the vector a. The partial-cut landowner’s discrete-choice problem is to choose the forest type s_j to leave on the ground as a seed source for new growth. The optimized value of the post-harvest (ph) land conditional on harvest method (h) is:

\begin{matrix} (1) & V_{t}^{ph | h} (s, a, c_{t}) = \max {\begin{matrix} V_{t} (s_{1}, a, c_{t}) \\ \begin{matrix} V_{t} (s_{2}, a, c_{t}) \\ ⋮ \end{matrix} \\ V_{t} (s_{S}, a, c_{t}) \end{matrix}} for h \in {CC, PC}, \end{matrix}

where S is the discrete number of different forest types that can physically grow on the land, and V_t(s_j, a, c_t) is the optimized present value of planting (or regenerating) the land with species s_j and which depends on period t climate conditions c_t. The vector of climate conditions could consist of all known and/or expected climate conditions as of period t, including potential future changes in climate that the landowner believes will occur. The age vector a consists of all zeros for bare land that was clear-cut (

h = CC

), and a potential mix of ages for land that was partial-cut (

h = PC

) and still has some standing trees.

Now consider a landowner of a stand with age vector a, whereby he can choose to clear-cut harvest, partial-cut harvest, or not cut and let his stand continue to grow. If the landowner harvests his land, he receives a net revenue from current-period harvest equal to $V_{t}^{h} (s, a) = P_{s} {vol}^{h} (a, s) - {HC}^{h}$ , where P_s is the unit price of forest type s, vol^h represents timber volume given choice of harvest method h where ${vol}^{CC} (a, s) > {vol}^{PC} (a, s)$ , and the variable HC^h represents harvest costs for harvest method h. The landowner’s forest management choice problem can be set up as jointly choosing harvest and replanting to maximize his land value function:

\begin{matrix} (2) & V_{t} (s, a, c_{t}) = max {\begin{matrix} V_{t}^{h = CC} (s, a) + ρ V_{t + 1}^{ph | CC} (s, 1, c_{t + 1}) \\ V_{t}^{h = PC} (s, a) + ρ V_{t + 1}^{ph | PC} (s, a^{'} + 1, c_{t + 1}) \\ ρ V_{t + 1}^{ph | NC} (s, a + 1, c_{t + 1}) \end{matrix}}, \end{matrix}

where

ρ = 1 / (1 + δ)

is a discount factor and δ is the discount rate. If the landowner harvests his land in either a clear-cut (CC) or partial-cut (PC), he receives a one-time net revenue from harvest

V_{t}^{h} (s, a)

and his subsequent post-harvest land value is determined by the solution to equation (1),

V_{t + 1}^{ph | h}

. The age vector a′ represents the period

t + 1

age vector for the trees that the landowner left standing in a partial-cut harvest. For the landowner who clear-cuts his land, all trees are of age 1 in period

t + 1

. If a landowner chooses not to harvest his land, then his post-harvest value function from choosing “no-cut” (NC) is

ρ V_{t + 1}^{ph | NC} (s, a + 1, c_{t + 1})

, which is affected by stand and price growth as well as the risk that the stand may be naturally disturbed by wildfire, insect damage, or diseases. The landowner chooses not to harvest his land when land value is maximized by leaving the stand to grow an additional period, that is,

V_{t} (s, a, c_{t}) = ρ V_{t + 1}^{ph | NC} (s, a + 1, c_{t + 1}) = ρ V_{t + 1} (s, a + 1, c_{t + 1})

Now consider the introduction of a carbon price, whereby the landowner receives a yearly subsidy of $P_{C} \cdot vol c$ , where P_C is the carbon price and volc is the carbon sequestered for each unit of timber added to the growing stock. If the landowner harvests his land, he is taxed at harvest by the amount of carbon released, and so the one-time net revenue from harvest $V_{t}^{h} (s, a)$ is augmented with a tax of $- P_{C} (1 - l) {vol c}^{h} (a, s)$ , where l is the fraction of harvested timber that continues to sequester carbon, and volc^h(a, s) is the volume of carbon from harvest method h (app. C). This is the setting introduced in van Kooten et al. (1995) and commonly used in the forest economics literature (Susaeta et al. 2014; Ekholm 2016). The effect of carbon pricing can be thought of as a carbon rent, which is the annualized discounted present value of the carbon sequestration benefits over all future rotations. Since the rate of carbon sequestered in volc is a function of the planted forest type s, a carbon price will change the replanting optimization in equation (1) and incentivize the landowner to replant the species that sequesters the most carbon. If climate change increases the volume of forest type s_j at each age a relative to every other species in the landowner’s choice set, and if carbon sequestered is proportional to the physical quantity of timber,² then a fixed carbon price will reinforce the effects of climate change in terms of raising the land value of planting s_j relative to the land value of planting the other forest types.³ Therefore, a carbon price affects an optimizing landowner’s choices associated with harvest timing, harvest method, and replanting (eq. [2]).

3. Empirical Econometric Framework

3.1. Specification of Nested Logit Model of Forest Management

In order to apply the forest management choices in equation (2) to empirical data, we require a framework that accounts for the fact that numerous drivers of the value function in equation (2) are observable to landowners but unobservable to empirical researchers. We integrate the basic theoretical setup above with a random utility interpretation of a nested logit model that accounts for observable and unobservable features of the management problem in (2). Our estimation structure explicitly embeds the solution to the discrete-choice replanting problem in equation (1) into the discrete-choice harvest problem in equation (2). As shown in figure 3, we divide the landowner’s forest management choice set into mutually exclusive harvest groups $h_{k} (k = 1, \dots, K)$ , each containing a post-harvest management/disturbance outcome $j (j = 1, \dots . J_{k})$ . For our application, we have $K = 3$ harvest groups (clear-cut, partial-cut, no-cut). If the landowner clear-cuts his land, then there are $J_{k} = 6$ forest types in which the land can be replanted. If the landowner partial-cuts his land, then there are $J_{k} = 6$ forest types that can be naturally regenerated by choosing which trees are left standing as a seed source. If the landowner chooses not to cut his trees, then there are $J_{k} = 2$ potential outcomes in which the stand can be naturally disturbed (e.g., fire, insects) or not.

**Figure 3.**
Nested structure of harvest and replanting decisions. Replanting choices 1 through 6 correspond to Douglas-fir, fir/spruce/mountain hemlock, hemlock/Sitka spruce, ponderosa pine, other softwood, and hardwood, respectively.
View Large Image Download PowerPoint

Combining our behavioral model in section 2 with Train’s (2009) decomposition of a nested logit model into two separate logit models, let landowner n’s value function associated with forest management action j in time t equal:

\begin{matrix} (3) & V_{njt} (s, a, c_{t}) = V_{nkt}^{h} (s, a) + V_{njt}^{ph | h} (s, a, c_{t}) + ε_{njt}, \end{matrix}

where

V_{njt}^{ph | h}

is unique to post-harvest outcome j, and

V_{nkt}^{h}

is unique to harvest choice k and common to all post-harvest outcomes in h_k, including k. The term ε_njt is observable to the landowner but not to the researcher and is assumed to be distributed generalized extreme value.

The primary assumption in section 2 is that landowner n chooses management action j in time t to maximize his land value function V_njt, and Train (2009) shows how this type of discrete-choice maximization problem generates an estimable probability that landowner n chooses management action j in time t as a product of two logit models, the probability of harvest action k multiplied by the probability of post-harvest outcome j conditional on choosing harvest k:

\begin{matrix} (4) & {Prob}_{njt} = {Prob}_{nkt} \cdot {Prob}_{njt | k} = \frac{exp (V_{nkt}^{h} + λ_{k} I_{nkt})}{\sum_{k = 1}^{K} exp (V_{nkt}^{h} + λ_{k} I_{nkt})} \cdot \frac{exp (V_{njt}^{ph | h} / λ_{k})}{\sum_{j = 1}^{J} exp (V_{njt}^{ph | h} / λ_{k})} . \end{matrix}

The term

I_{nkt} = \ln \sum_{j = 1}^{J} exp (V_{njt}^{ph | h} / λ_{k})

is known as an inclusive value for nest k, and λ_k is a parameter to be estimated. Importantly, the probability of harvest choice k is a function of the inclusive value and, hence, the probability of harvest is necessarily affected by the drivers of all post-harvest outcomes. Therefore, the nested logit model incorporates a key point from section 2 in that the landowner’s optimal replanting choice from equation (1) is structurally embedded into the landowner’s optimal harvesting choice from equation (2).

To specify the empirical model with observable data, we begin with the lower nest describing the replanting choice conditional on the landowner having clear-cut or partial-cut his land. In general, the post-harvest value function $V_{njt}^{ph | h}$ depends on a potentially complicated function of the landowner’s expectations of future prices, tree growth, and climate change and could include multiple anticipated switches between planted forest types. Guo and Costello (2013) provide an example of numerically estimating $V_{njt}^{ph | h}$ under the assumption that landowners dynamically optimize management under an anticipatory expectation of how climate change will affect growth—and, hence, profitability—from replanting different forest types. However, evidence from extension research in the Pacific Northwest suggests that forest landowners are currently not accounting for anticipated future climate change in their management actions (Grotta et al. 2013); plus, forest landowners and appraisers in our study region are specifically trained to estimate forestland values using a static expectations Faustmann formula.⁴ Finally, large uncertainties in downscaled climate forecasting mean that “it is likely that a great deal of climate adaptation will be reactive rather than anticipatory” (Massetti and Mendelsohn 2018, 335).

In choosing which forest type to replant, landowners are assumed to compare the expected rents that their land would generate when planted with different species. We specify $V_{njt}^{ph | h}$ for replanting (bottom left and center of fig. 3) as a reduced-form function of the average per-acre value function for planting species s_j in region r that contains plot n, annualized as a rent: ${\bar{rent}}_{r (n) s_{j} t}$ , where the upper bar notation indicates a regional average. Regional rents are a function of forest growth, forest type specific prices, and site productivity. To compute expected rental values, the bare land values are first calculated for six forest types, seven site productivity classes,⁵ and 18 price regions across the three states in our study region.⁶ Using the FIA data, we empirically fit separate yield curves for each forest type and site productivity classes by price region, which are then used to compute approximate Faustmann optimal rotation lengths. Rents are then imputed using a 5% discount rate (app. C).

A forest owner can select a forest type to replant from the following six types: (1) Douglas-fir, (2) fir/spruce/mountain hemlock, (3) hemlock/Sitka spruce, (4) ponderosa pine, (5) other softwoods,⁷ and (6) hardwoods.⁸ One of the alternative specific constants is set to zero for identification. We use observable variation in climate within each region to infer the relationship between climate and replanting choice, using the revealed behavior of replanting choice. We do this by including interaction terms between our regional rents and the more downscaled climate variables in the replanting equation. We specify the nested logit model for the replanting nest as the following function:

\begin{matrix} (5) & \begin{matrix} V_{njt}^{ph | h} = β_{0 j}^{ph} + γ^{ph} {\bar{rent}}_{r (n) s_{j} t} + β_{1 j}^{ph} {\bar{rent}}_{r (n) s_{j} t} \cdot c_{nt} + β_{2 j}^{ph} {elev}_{n} + β_{3 j}^{ph} {\bar{rent}}_{r (n) s_{j} t} \cdot c_{nt + 30} \\ for ph \in {replant | clear-cut, regenerate | partial-cut}, \end{matrix} \end{matrix}

where

{\bar{rent}}_{r (n) s_{j} t}

and c_nt are rent and downscaled climate variables representing historical long-run averages. Interactions between

{\bar{rent}}_{r (n) s_{j} t}

and c_nt describe observable rent deviations between plot n and the regional average; elev_n is the elevation of plot n, and

c_{nt + 30}

is the projected change in mean temperature based on the forecasted climate 30 years into the future. Our inclusion of Faustmann rents into the specification of

V_{njt}^{ph}

is meant to provide a reasonable and observable index for how prices, timber growth, and approximate expected rotation times influence the post-harvest land value function, recognizing that landowners have many unobservables (expectations, management skills, etc.) that also affect the value function and are embedded in the logit unobservable, ε_njt. Similar to the logic of Severen et al. (2018), our reduced-form approach to including

c_{nt + 30}

in equation (5) provides a simple test for whether current economic decisions reflect a downscaled climate forecast in an anticipatory fashion.

The parameters for the interaction terms in (5) are specific to each replanting choice, thus revealing the relationship between climate and the expected value of forestland associated with each species replanted. Figure 4 graphically illustrates the intuition from this revealed preference approach. The horizontal axis represents a climate measure (c) such as temperature. In figure 4A, the climate is represented by a probability distribution, or the likelihood that each temperature occurs. Figure 4B represents the relationship between the value of forestland and climate. At $c \leq \bar{C}$ , planting Douglas-fir is optimal since the value of the land is highest in that use, while planting (regenerating) hardwoods is optimal at $c > \bar{C}$ . Since we do not observe plot-level value functions of forestland, we instead use revealed planting choices at the current climate distribution f1 to estimate parameters in (5) which then generate the probability of replanting a forest type as a function of climate c. If the climate distribution shifts from f1 to f2 in figure 4, then our estimated model would predict an increase in probability that landowners will plant hardwoods with a corresponding decrease in the probability of planting Douglas-fir (i.e., $PD 2 < PD 1$ in fig. 4).

**Figure 4.**
Conceptual relationship between climate, the value of forestland, and the revealed probability of replanting under a climate distribution.
View Large Image Download PowerPoint

We account for the risk of natural disturbance in estimation through the “no-cut” nest (bottom right of fig. 3), where landowners refrain from cutting their timber in exchange for letting the trees grow an additional period. By choosing not to cut, the landowner leaves the stand at risk to the binary outcomes of natural disturbance or no disturbance. Since fire risk influences the landowner’s harvest decision (Reed 1984), we jointly estimate drivers of disturbance and harvest decisions by specifying the nested logit model for the lower “no-cut” nest as a binary model:

\begin{matrix} (6) & \begin{matrix} V_{njt}^{ph | h} = ω_{0}^{ph} + ω_{1}^{ph} {priv}_{n} + ω_{2}^{ph} {elev}_{n} + ω_{3}^{ph} {species}_{nt} + ω_{4}^{ph} {vol}_{nt} + ω_{5}^{ph} c_{nt} + ω_{6}^{ph} {state}_{n} \\ for ph \in {disturbance event | no-cut} . \end{matrix} \end{matrix}

The independent variables include an ownership dummy indicating private or state ownership (priv_n), elevation (elev_n), forest type dummy variables indicating the forest type (species_n), the current timber volume (vol_nt), a state dummy (state_n), and a vector of climate variables (c_nt). Climate variables such as precipitation directly affect nature’s ability to suppress fires, while other climate variables such as minimum winter temperatures can affect the susceptibility of certain trees to damage. We do not have data on past fire management activities for each plot (e.g., thinning, controlled burns, etc.), which end up in the econometric unobservable. Our current specification assumes that unobserved fire management activities are uncorrelated with the independent variables in (6) (see app. I for robustness checks).

Now consider the upper nest in figure 3, whereby the forest landowner makes the harvest decision by choosing whether to clear-cut, partial-cut, or not cut his stand of trees. Following Provencher (1997), we specify the observable components specific to the net revenue from harvest method h as:

\begin{matrix} (7) & \begin{matrix} V_{nkt}^{h} = α_{0 k} + α_{1 k} P_{{ns}_{j} t} \cdot {vol}_{nk (s_{j}) t}^{h} \\ for h \in {clear-cut, partial-cut} . \end{matrix} \end{matrix}

And we specify the observable components specific to the decision not to cut and let the stand grow as a function of expected changes in revenue:

\begin{matrix} (8) & \begin{matrix} V_{nkt}^{h} = α_{1 k} P_{{ns}_{j} t} \cdot Δ {vol}_{nk (s_{j}) t} \\ for h \in {no-cut} . \end{matrix} \end{matrix}

In this specification, we use observable time t timber prices for forest type s_j from the region that contains plot n, and multiplied by the observable species s_j timber volume for management choice k as a representation of the one-time revenue that the landowner would receive from picking harvest choice k. Since clear cutting necessarily entails harvesting more volume than partial cutting,⁹ the volume variable is indexed by harvest choice k. The post-harvest value function from not cutting the land—

V_{t + 1}^{ph | NC} (s, a + 1, c_{t + 1})

in equation (2)—is affected by the marginal benefit of waiting to cut, which is the change in revenue that could be received by allowing the stand to grow an additional period.¹⁰

Finally, the nested logit structure embeds the inclusive value I_nkt of the lower post-harvest nests into the upper harvest nest of the estimated probabilities in equation (4). For the two harvest alternatives (clear-cut and partial-cut), I_nkt approximates the optimized post-harvest land value associated with picking the forest type to replant (Hartman 1988; Train 2009), which is a direct measure of the solution to equation (1).¹¹ With the inclusive value from each nest, climate implicitly affects the harvest decision and so this empirical framework allows the climate to affect adaptation on the extensive margin (choosing which forest type to plant) and on the intensive margin (altering the harvest time). The alternative specific constant of “no-cut” is normalized to zero.

The parameters defining the probabilities of harvest, disturbance, and replanting choices are simultaneously estimated with maximum likelihood techniques using original MATLAB code. The log-likelihood function is:

\begin{matrix} (9) & LL (α, β, γ, λ) = \sum_{n} \sum_{j} y_{nj} \ln {Prob}_{njt}, \end{matrix}

where y_nj equals one if landowner n chooses management j. We index some variables in the model with time t to represent that different plots are observed at different points in time, and so different plots have variables measured at different points in time. This is a pooled rather than a panel data model.¹² Finally, we weight each plot’s likelihood by the expansion factor assigned in the FIA database, where the expansion factor represents the sampling intensity associated with the sample plots.

If a plot is naturally disturbed with wildfire, we use historical spatial-temporal data on wildlife burn severity to separately estimate a burn severity index as a function of the same climate and other plot-level drivers of natural disturbance from equation (6):

\begin{matrix} (10) & {SV}_{n} = G ({priv}_{n}, {elev}_{n}, {species}_{nt}, {vol}_{nt}, c_{nt}, {state}_{n}; ϑ), \end{matrix}

where SV_n is the most dominant burn severity (1: unburned to low, 2: low, 3: moderate, or 4: high) that has occurred in the last 10 years within a 2 km radius around each plot. The vector of parameters ϑ are estimated as an ordered logit model (see app. E) using historical burn severity data from 2001 to 2014. The parameters defining the probabilities of burn severity are estimated outside of the nested logit framework, as coupling an ordered logit model to the nested logit model is computationally challenging and brings up convergence concerns. Estimating burn severity parameters separately means that we are not accounting for how variation in the burn severity level will affect land value and management. Instead, we use the projected probabilities of severity in the simulation (see sec. 5) to adjust the stand volume, age, and incremental growth.

3.2. Data

We have plot-level data from FIA for over 6,800 forested plots with variables representing forest type, site quality, tree growth, elevation, an indication of recent harvest, and an indication of recent natural disturbance. We combine the plot-level forest management data with downscaled climate data and regional timber prices that vary across forest types and site quality. The study area of Oregon, Washington, and California has a substantial portion of its landscape dedicated to commercial forest production (30%, 44%, and 45% of nonfederal rural lands are forested in California, Oregon, and Washington, respectively), including some of the most productive forests in the world, and has considerable climate variation and corresponding variation in forest types (fig. 5A). Due to a methodological change in data collection, the FIA is only available since 2001. Since we do not observe multiple harvests on the same plot, a complete set of panel data does not exist.¹³

**Figure 5.**
Nonfederal FIA forest plots on US West Coast. A, Current distribution of forest types. B, Projected climate change. “Much warmer” and “much drier” indicate 4°C or more increase in temperature and larger than 50 mm precipitation decrease, respectively.
View Large Image Download PowerPoint

Each plot is designed to cover 1 acre and is randomly sampled. Plot locations are slightly “fuzzed” for confidentiality in that each plot’s true location is within at most 1 mile (1.6 km) of its stated location. Using a national standard for field measurements for a wide range of site attributes, the FIA presents the most detailed plot-level data available for our purposes.¹⁴ The key dependent variables for econometric analysis include a qualitative indicator of harvest on an FIA plot (clear-cut, partial-cut, no-cut), the forest type replanted upon harvest, and the presence of a natural disturbance (e.g., fire, insect damage, etc.). The plot-level attributes used include forest type, stand volume, incremental volume growth, owner type, stand age, and elevation. We assigned each plot to a specific timber price region for different forest types. Historical timber prices for different grades are drawn from records available by state-level agencies. Since not all forest types can physically grow across the entire region, each plot is assigned a choice set based on the “plant viability scores” developed by the USDA Forest Service that reflects the likelihood that the climate at a given location would be suitable for each species (Crookston et al. 2010).¹⁵

The climate variables observable to the landowners are total precipitation and mean temperature during the growing season,¹⁶ the maximum temperature in the warmest month (August), and minimum temperature in the coldest month (December). These variables have been found to be some of the most influential variables that affect the growth of trees (Rehfeldt et al. 2014). Plot-level climate data are based on normal monthly data from the Parameter-elevation Regressions on Independent Slopes Model (PRISM) at 800 m resolution over a 30-year period between 1981 and 2010. The climate projection that we use is of high spatial resolution (1 km) and is easily available to the public online.¹⁷

Appendix F presents a full list of data sources and summary statistics of the data, but a few highlights are covered here. First, 9.3% (5.6%) of forest plots were clear-cut (partial-cut) over the 10-year sample period for each plot. Second, on a per-unit basis, Douglas-fir logs were the most commercially valuable forest type ($458.50/thousand board feet) and comprise about 50% of total harvested volume, and hardwoods are the least valuable ($254.49/thousand board feet). Third, Douglas-fir and hemlock/Sitka spruce are more commonly harvested by clear-cut in the wetter region west of the major mountain ranges, while ponderosa pine is more commonly harvested by partial-cut in the drier region east of the major mountain ranges. Finally, fir/spruce forest types are most commonly found in the mountains at high elevation while hardwoods are most commonly found in the valleys at a lower elevation.

4. Econometric Estimation Results

The full set of nested logit parameter estimates and partial effects for key variables are presented in appendix G. Parameter estimates (table G1) show that replanting decisions and natural disturbance are significantly affected ( $p < .05$ ) by a variety of climate variables and generally conform to expectations and yield four general conclusions. First, in the lower replanting nest, the estimated coefficient on the rent variable is positive and highly significant ( $p < .01$ ) as expected—landowners are more likely to replant forest types that are more profitable. Second, the rent-climate interaction parameters are jointly significant in the replanting nest ( $p < .01$ ) but individual parameters are highly variable in magnitude and statistical significance across the different replanting choices. About 40% of the individual climate parameters are significantly different from zero. Third, natural disturbance is more likely on high elevation, large volume, and dry plots that experience cold winter temperatures and that are hardwood or “other softwood” forest type. Fourth, in the upper harvest nest, the estimated coefficients are significantly different from zero ( $p < .05$ ) and consistent with standard comparative statics from harvesting models—a timber harvest is more likely when the marginal costs of waiting to harvest are high (i.e., harvest revenue; inclusive value proxy for optimized bare land value), and less likely when the marginal benefits of waiting to harvest are high (i.e., growth in harvest revenue).

A notable partial effect (table G2) is that the probability of replanting Douglas-fir declines by approximately 30 percentage points given a 3°C increase in mean temperature ( $p < .01$ ) for the prime growing regions of western Oregon and Washington—a direct estimate of $PD 2 - PD 1$ from figure 4. We find limited evidence that our variables representing future 30-year forecasts of mean temperature significantly affect replanting. The partial effect of a 1°C increase in forecasted future mean temperature has no significant effect on the probability of replanting all forest types except ponderosa pine, where the partial effect is −8.2% and only marginally significant ( $p < .1$ ). Thus, we find strong evidence that current replanting decisions react to current climate, but limited evidence that current replanting decisions respond to anticipated forecasts of future climate.

The partial effect of rent on the probability of replanting each forest type is the primary mechanism that we use to simulate the impacts of a carbon rental payment on forest composition. Given our use of interactions between regional rents and downscaled climate variables in equation (5), the partial effects of replanting with respect to rent depend on climate. Figure 6 shows the partial effect of rent on replanting two main forest types in the region—Douglas-fir and hardwoods. As seen in figure 6A and figure 6B, the average partial effects of rent on replanting Douglas-fir fall with climate change, while the average partial effects of rent on replanting hardwoods are roughly the same with climate change. By basing estimation off landowners’ revealed preferences for replanting forest types at locations with different current climates, the estimated partial effects pick up the declining productivity of Douglas-fir with the warming climate that we hypothesized in the conceptual relationship from figure 4. As a check on this finding, we compare our estimated partial effects to natural science projections of the biophysical viability of Douglas-fir and hardwoods from Crookston et al. (2010) where higher viability scores indicate higher viability.¹⁸ Figure 6C and figure 6D show average viability scores for Douglas-fir and hardwood forest types for our study region as a function of the climate projections. Both the econometric model and the natural science model produce results indicating strong declines in the productivity of Douglas-fir with the warmer climate. Hardwoods become somewhat less viable in the natural science model, while the econometric results show no strong decline in productivity. The similar impact of climate on both the econometric partial effects and the biophysical viability scores provides a qualitative convergent validity check.

**Figure 6.**
Partial effects of $10/acre rent increase on replanting probabilities for Douglas-fir (A) and hardwoods (B) and average plant viability scores (from Crookston et al. 2010) for Douglas-fir (C) and hardwoods (D). Dashed lines in C and D indicate range where species has little chance to be biophysically viable. Below the dashed lines the species has almost no chance to be viable.
View Large Image Download PowerPoint

Finally, appendix I presents results from nine sets of robustness checks of the econometric model, which indicate that the model is quite robust to alternative specifications.

5. Landscape Simulation Analysis

On their own, the parameter and marginal effect estimates from the econometric model do not give a full picture of the potential landscape shifts under climate change. Replanting does not occur continuously, but only after infrequent harvest events (∼15% probability of a plot being harvested over 10 years). The projected climate changes are also heterogeneous across plots. Since multiple climate variables will shift in a spatially heterogeneous fashion simultaneously, simulating the combined effects when all climate variables shift at the same time will render more informative results. A landscape simulation will need to take these factors into consideration in order to examine the effect of a path of climate changes on landscape change. As described in section 1, we use our econometrically estimated forest management and natural disturbance probabilities as a set of decision rules to simulate multiple realizations of changes in the landscape of forest types under a changing climate and a carbon price policy.

5.1. Simulated Climate Change and Carbon-Pricing Scenarios

We simulate three scenarios of landscape change. First, the baseline scenario assumes no climate change or carbon price and simply extends the observed forest management practices from 2001 to 2014 into the future. Second, the climate change only scenario assumes that future climate regimes are derived from the US National Center for Atmospheric Research Community Climate System Model (CCSM) 4.¹⁹ Out of the available scenarios included in the model, we chose the RCP 8.5 scenario, as current CO₂ emission rates are closely tracking this pathway (Sanford et al. 2014; McKenney et al. 2015). Under the RCP 8.5 scenario, the majority of premier private forestland in the west side of the Cascades, the area shown in brown in figure 5b, is expected to become warmer and drier. The average temperature is expected to increase by 4.35°C by 2100 on average in the study region.

In the carbon-pricing scenario, we use the RCP 8.5 climate projection and add a hypothetical carbon-pricing scheme. Our carbon-pricing scheme starts in 2020 and assumes landowners receive rental payments for the amount of carbon sequestered in their forests. We assume that a carbon price starts at $15/ton in 2020, rises to $50 in 2050, and again to $80 in 2080. To determine carbon payment values to landowners, we estimate regional average carbon yield curves from the FIA data, where tons of carbon sequestered is expressed as a function of stand age. The carbon yield curves vary by forest type, region, and site class. Given a set of timber prices, carbon prices, and yield functions for each region r, we translate the carbon price to a regional average per-acre rental payment for each plot n that is in region r and is planted with species $s_{j} ({\bar{rent}}_{r (n) s_{j}}^{TC})$ , where the “TC” superscript indicates that landowners are being paid for both timber and carbon sequestration. We calculate ${\bar{rent}}_{r (n) s_{j}}^{TC}$ using the approach of van Kooten et al. (1995), which subsidizes landowners for carbon sequestered through tree growth while taxing landowners for carbon released at harvest.²⁰ We calculate ${\bar{rent}}_{r (n) s_{j}}^{TC}$ under the assumption that landowners solve for the rotation length that maximizes the present value of timber and carbon benefits. The carbon price scenario is implemented by replacing the variable ${\bar{rent}}_{r (n) s_{j} t}$ (timber rents only) in the lower replanting nests in equation (5) with ${\bar{rent}}_{r (n) s_{j}}^{TC}$ (timber and carbon rents). This approach assumes that a landowner is indifferent between $1 in carbon rent and $1 in timber rent. The value of ${\bar{rent}}_{r (n) s_{j}}^{TC}$ reflects the regional average timber and carbon rent at today’s climate distribution. The effects of climate change on timber and carbon rents are accounted for by the econometric parameters in (5), which translates any given combination of ${\bar{rent}}_{r (n) s_{j}}^{TC}$ and climate c_nt for each plot n to forest management probabilities. As discussed in section 4, our econometric results imply that the average Douglas-fir plot is becoming less productive under climate change in western Oregon and Washington, and thus the econometric model implicitly lowers the carbon sequestration productivity of the average Douglas-fir plot under the assumed RCP 8.5 climate change path.

Differences between the climate change only scenario and the baseline give us the impact of climate change on the resulting landscape. Differences between the carbon-pricing scenario and the climate change only scenario give us the impact of carbon pricing alone, while the difference between the carbon-pricing scenario and the baseline gives us the combined impact of climate change and carbon pricing.

Future timber prices are projected to increase by 0.4% and 0.5% per year under the climate change scenario and baseline scenario, respectively, using results from Sohngen and Tian’s (2016) global timber model.²¹ To adjust the forest growth functions to projected climate change, we adapt the net primary productivity (NPP) projection for Douglas-fir developed by Nicholas Coops and Richard Waring in their 3-PG process model (Coops et al. 2010). As shown in appendix J, NPP is expected to decline relative to the 1990 level in most of the west side of the study region and increase in most of the east side. For each plot, incremental tree growth at each future time step is adjusted by multiplying the incremental tree growth with the downscaled NPP at each location and at each time step relative to NPP in the base year (e.g., if NPP in 2020 is 30 and NPP in 1990 is 32, relative NPP is $30 / 32 = 0.93$ ). Regional timber rents are also adjusted to reflect the biophysical effects of climate on NPP.

5.2. Summary of Key Assumptions That Drive the Landscape Simulation

Our simulation uses the estimated plot-level management probabilities to project changes in the US Pacific coast forest stock by simulating and aggregating plot-level management and natural disturbance outcomes in response to a number of exogenously changing variables. We assume that climate exogenously changes under RCP 8.5 and that global timber prices and the NPP of forests change in response to climate change as estimated by Sohngen and Tian (prices) and Coops et al. (NPP). Within the simulation, the estimated management/disturbance probabilities from our discrete-choice model determine how the timing and intensity of harvest, the forest type to replant, and natural disturbance respond to the exogenously changing variables within multiple 10-year time steps. Our estimated econometric model imposes an assumption that landowners have static expectations of exogenous factors—in time t, landowners assume that the level of the exogenous variables (climate, prices, NPP) will continue indefinitely at period t levels.²² Since the simulation operates at 10-year time steps, then landowners respond to changes in the exogenous factors and re-optimize their management decisions every 10 years. For example, the estimated rental value for Douglas-fir that determines the probability of replanting Douglas-fir in 2060 is a function of the constant levels of all exogenous variables that occur in 2060 (climate, prices, NPP), which are different than the exogenous variable levels that were used to calculate the same Douglas-fir rental value in 2020. Our simulation approach implicitly implements Massetti and Mendelsohn’s (2018) argument that climate adaptation will be reactive rather than anticipatory, with people adjusting to the changes in climate that they actually observe. Previous large-scale econometric landscape simulation exercises have operated under similar assumptions about expectations (e.g., Lubowski et al. 2006; Lawler et al. 2014). In contrast, market simulation analyses of forestry assume that markets perfectly anticipate future climate change (e.g., Sohngen and Mendelsohn 1998; Lee and Lyon 2004).

5.3. Landscape Changes under Baseline, Climate Change, and Carbon-Pricing Scenarios

For each scenario, we calculate the average share of nonfederal forestland in each forest type and for each time step, averaged over the 1,000 Monte Carlo simulated landscape outcomes. Figure 7 presents our projection of each forest type’s composition in California, Oregon, and Washington for the baseline, climate change, and carbon-pricing scenarios.

**Figure 7.**
Time path of each forest type’s total landscape share under baseline scenario with no climate change (*left*), climate change only scenario (*middle*), and climate change and carbon-pricing scenario (*right*). DF: Douglas-fir, FIR: fir/spruce, HEM: hemlock/Sitka spruce, PP: ponderosa pine, OS: other softwood, and HW: hardwood.
View Large Image Download PowerPoint

Examining the three scenarios provides three conclusions. First, climate change induces a reduction in the share of nonfederal forestland in the commercially dominant Douglas-fir type in Oregon and Washington and induces an increase in the share of ponderosa pine (for California) and hardwoods (all states).²³ Second, carbon pricing reinforces these shifts. Third, the extent of landscape change is subtle since replanting harvested land—thus replacement of a landscape’s forest type—occurs gradually on a fraction of the landscape following infrequent harvest events. The seemingly subtle landscape change is masking a change in replanting preference that is happening much more drastically. For example, figure 8 shows that while 50% of all harvested land in Oregon is currently replanted with Douglas-fir, only 16% of harvested land in Oregon is projected to be replanted to Douglas-fir by 2090 under climate change and carbon pricing. This is a sizable shift in the preference for the most commercially dominant tree in the Pacific Northwest. In contrast, hardwoods currently comprise 12% of replanting/regeneration on harvested land in Oregon, but that share rises to 53% by 2090 under climate change and carbon pricing.

**Figure 8.**
Time path of each forest type’s total replanting share under climate change only scenario (*left*) and carbon-pricing scenario (*right*). The kinks in the chart are due to the change in the replanting choice set that is assigned to each plot. The choice set is based on the plant viability scores that reflect the likelihood that the climate at a given location would be suitable for each species.
View Large Image Download PowerPoint

Figure 9 highlights regional differences for the two forest types that experience the largest changes—Douglas-fir and hardwoods. We compare total landscape change over the whole period as well as change on harvested plots only in the final period (2090–2100). Compared to the landscape change, landowners dramatically increase the replanting of hardwood species on harvested land in 2090 at the expense of Douglas-fir in Oregon and Washington in particular—a region that is mostly predicted to become warmer and drier under the climate scenario we use. Both climate change and carbon pricing induce the shift from Douglas-fir to hardwoods. As discussed in section 4, Douglas-fir gets less productive under climate change, as reflected by the declining estimated partial effect of rent on replanting Douglas-fir with the warming climate. Since most Douglas-fir plots in our region are artificially regenerated by planting, while most hardwoods are naturally regenerated following harvest, a switch from Douglas-fir to hardwoods could be interpreted as landowners abandoning the management-intensive form of artificial regeneration (replanting) of Douglas-fir in favor of less management-intensive natural regeneration of hardwoods.

5.4. Probability of Changing Forest Types (Extensive Margin Adaptation)

The above simulation results are averages across 1,000 simulated landscape changes. The speed of landscape change is driven by the magnitude of harvest and natural disturbance in each 10-year period, while the direction of change is driven by the replanting/regeneration choices. To exploit further information in the estimated probabilities of forest management, we calculate the effects of climate change and carbon prices on the probability that landowners change from their initial forest types to a different forest type by 2100. We interpret this measure as a probability of adaptation on the extensive margin. For plot n that begins today in forest type s_j, the simulations generate 1,000 realizations of plot n’s forest type by the year 2100. The probability of extensive margin adaptation for plot n is defined as ${Prob}_{n} (adapt) = [1 - {Prob}_{n} (s_{j, 2100} = s_{j, 2010})]$ , which is calculated as the proportion of the 1,000 simulations in which plot n ends the year 2100 in a different forest type than it began in 2010. Subtracting the baseline calculation of Prob_n(adapt) from the climate change calculation of Prob_n(adapt) gives the discrete effect of a 90-year climate path on the probability of adaptation. Similarly, subtracting the climate change calculation of Prob_n(adapt) from the carbon-pricing scenario version of Prob_n(adapt) gives the discrete effect of a 90-year carbon price path on the probability of adaptation. Table 1 presents results for western Oregon and Washington as an example, differentiated by initial forest type and final forest type. A full table for all regions is included in appendix K.

Table 1.

Discrete Effect of a 90-Year Climate and Carbon Price Path on the Probability of Extensive Margin Adaptation by Switching from Forest Type and Switching to Forest Type (Western Oregon and Western Washington as an Example)

Region and Initial Forest Type (Switching From)	Effect of Climate Change	Effect of Carbon Price	Final Forest Type (Switching To)	Effect of Climate Change	Effect of Carbon Price
West Oregon:
Douglas-fir	10.6%	6.9%	Douglas-fir	−5.5%	−2.0%
Fir/spruce	−6.7%	−2.1%	Fir/spruce	2.5%	.4%
Hemlock/Sitka	−3.6%	.2%	Hemlock/Sitka	−3.8%	−.3%
Ponderosa pine	−8.8%	2.4%	Ponderosa pine	1.3%	.7%
Other softwood	−11.4%	.2%	Other softwood	.6%	−.1%
Hardwood	−16.5%	−4.3%	Hardwood	6.8%	4.6%
West Washington:
Douglas-fir	1.9%	4.2%	Douglas-fir	−5.6%	−2.1%
Fir/spruce	−10.2%	−1.0%	Fir/spruce	.3%	−.2%
Hemlock/Sitka	−2.7%	1.2%	Hemlock/Sitka	−4.1%	−.9%
Ponderosa pine	NA	NA	Ponderosa pine	.0%	.0%
Other softwood	−11.7%	−.3%	Other softwood	1.6%	1.3%
Hardwood	−15.1%	−3.8%	Hardwood	3.9%	3.2%

View Table Image

Table 1 shows that the discrete effect of a 90-year path of climate change and carbon pricing varies across regions and initial forest types. For example, a Douglas-fir landowner in western Oregon is 10.6% more likely to adapt away from Douglas-fir on the extensive margin by the year 2100 under climate change. In contrast, the current owner of a hardwoods plot in western Oregon is 16.5% less likely to adapt to another forest type by 2100 under climate change. A key result from table 1 is that both a carbon price and climate change increase the probability that Douglas-fir owners in the prime-growing regions of western Oregon and Washington adapt to another forest type by the year 2100. Looking at the final forest types that the landowners are switching to, we find that the 90-year path of a carbon price has an effect similar to climate change—especially favoring hardwoods at the expense of Douglas-fir.

6. Discussion

The purpose of this paper is to quantify the effect of climate change and carbon pricing on the adaptation behaviors of private forest owners. While much attention focuses on climate change as a source of damage to human well-being, we focus on the fact that climate change is also an important input to many decisions, including forest management, and that adaptive decisions can induce changes to natural resource stocks that can cause changes in human well-being. Adaptation behavior in forestry affects landscapes directly, leading to a change in forest composition, which can then alter the flow of ecosystem services from forests. In order to understand the impact of climate change on landscape outcomes, we need to account for how landscape change is affected by deliberate adaptive decisions in response to climate, which in the case of forestry is also affected by policy interventions that aim to mitigate climate change. Our paper uses empirical evidence and simulation techniques to examine how climate adaptation, mitigation, and natural resources interact through land-use change. An important finding is that along the US Pacific coast, landowners adapt to climate change by gradually shifting out of their current dominant species choice of Douglas-fir to species more suitable for the future climate, notably hardwoods and ponderosa pine.

Our findings also show that carbon pricing accelerates forest landowners’ adaptation behaviors to climate change, reinforcing the forest landscape shift away from coniferous Douglas-fir forest types to other types such as hardwoods. The carbon price rewards the climate-induced relative shift in productivity away from Douglas-fir to hardwoods. A clear shift that involves two distinctively different forest types indicates that we should likely expect habitat losses for wildlife species that are specialized to Douglas-fir habitat, and future habitat gains for wildlife species that are specialized to hardwood forests. State wildlife agencies along the US Pacific coast recognize many wildlife species of conservation concern that are specialized to both coniferous forests²⁴ and hardwood forests,²⁵ and these species of conservation concern are sensitive to changes in private forests. Future research could link our projections of landscape change under climate change and carbon pricing with more detailed assessments of the spatial pattern and range of habitats for wildlife of conservation concern. Nevertheless, our results suggest that carbon pricing generates trade-offs between choices that are privately optimal to landowners against nonmarket changes in forest ecosystem services that may generate social costs to nonlandowners. This unintended effect of carbon pricing can lead to wildlife and biodiversity outcomes that have the potential to conflict with current policies focused on conserving existing wildlife habitat (e.g., the US Endangered Species Act). Our empirical approach could be extended to other regions to examine whether we would face similar trade-offs between private adaptation behavior and land-use changes within forestry.

There are caveats and areas in which future research could provide improvements to our modeling framework. First, a caveat associated with our results is that we are not able to differentially adjust timber prices for the forest types in response to the supply changes that we simulate. Ideally, we would have demand elasticity estimates for each forest type, which would allow us to endogenously translate any given supply shock in one forest type on the equilibrium price of that forest type. Second, future research could explore different structures of carbon benefits payment, especially a payment scheme that pays only for additional carbon sequestration. Third, our approach assumes that landowners reactively adapt to climate change rather than anticipate future climate change. While we provide limited econometric evidence in support of this assumption, future research could examine how much results would differ if landowners anticipated future climate changes in their management choices. Finally, it would be useful to include nonforest land uses such as pasture or range land for plots that are unable to grow any tree species in the future.

Carbon pricing could create local winners and losers with respect to those that consume ecosystem services from forests. Our simulation results show that climate change would lead to a lower present value of future harvest revenues for the average private forest owner in the next 90 years, though the carbon-pricing program we simulate would offset the loss. For example, the average present value of the stream of harvest and carbon payment revenues ($/acre) under the baseline, climate change scenario, and carbon-pricing scenario is $2,463, $2,197, and $2,485 (including the carbon subsidy and harvest penalty), respectively. These present values, which can be considered as dividend payments for two forest ecosystem services—timber and carbon sequestration—indicate that private landowners benefit from carbon pricing, but at the expense of other nonmarket ecosystem services that may be reduced by land-use changes within forestry. Relative to the baseline, Douglas-fir forests shrink by 13% (1.7 million acres) and by 21% (2.8 million acres) under the climate change and carbon-pricing scenarios, respectively. Dense Douglas-fir forests in the western portion of the study region lose the most acreage in our modeling results, which implies that there may be substantial regional heterogeneity in the localized impact of climate change and carbon pricing on habitat and other ecosystem services. Private forest landowners gain from carbon pricing, but nonlandowners would bear costs in the form of accelerating changes to wildlife habitat. In a second-best world of nonexistent markets for most ecosystem services such as wildlife habitat, there will be benefits from policy coordination between multiple second-best policies aimed at different goals, as well as determining what environmental benefits should be included when estimating the social benefits from policy implementation. Future research should examine the possibility of combining a carbon price with further policy incentives aimed at reducing negative externalities resulting from the landscape changes that come from private adaptation to a carbon price. “Think globally, act locally” resonates in the question of how to achieve mitigating global externalities when the mitigation options also generate local externalities.

Notes

Yukiko Hashida (corresponding author) is at the University of Georgia, Department of Agricultural and Applied Economics ([email protected]). David J. Lewis is at Oregon State University, Department of Applied Economics ([email protected]). Funding support from the USDA Forest Service Pacific Northwest Research Station (14-JV-11261955-059) and USDA National Institute for Food and Agriculture is gratefully acknowledged. We thank Darius Adams, Andrew Gray, Jeff Kline, David Kling, Christian Langpap, Brent Sohngen, and Eli Fenichel, as well as participants at the Oregon Resource and Environmental Economics Workshop, the 2016 Association of Environmental and Resource Economists summer conference, and seminar participants at Oregon State University, Landcare Research, Yale University, and the University of California, Berkeley, for useful comments. We thank two anonymous reviewers and the editor, Joshua Abbott, for multiple suggestions that greatly improved the paper.

1. We will mostly use the term “replant” to indicate how the landowner facilitates new tree growth on harvested land. Landowners in the Pacific states of the United States regenerate new growth through either management-intensive replanting (also known as artificial regeneration) or through less management-intensive natural regeneration. For example, most Douglas-fir stands are replanted while most hardwood and “other softwoods” stands are naturally regenerated. Other forest types are a mix of replanting and natural regeneration. See app. A for a breakdown of artificial regeneration in the study area.

2. We use the FIA data for carbon in the above-ground portion of the tree, which is derived by FIA crews as the sum of above-ground biomass estimates multiplied by 0.5.

3. Our assumption is that the land value function for tree species s_j is nondecreasing when the underlying tree growth parameters increase. Ceteris paribus, more tree growth is economically valuable.

4. For example, Oregon State University Forestry Extension experts teach valuation techniques for small woodland owners using a textbook Faustmann formula.

5. The site productivity class ranges from 1 to 7, where 1 indicates the most productive plot. This is a classification of forest land in terms of inherent capacity to grow crops of industrial wood expressed in cubic feet/acre/year.

6. There are four subregions in Washington, five in Oregon, and nine in California, each corresponding to a price region for which state agencies report regional timber prices.

7. Other softwoods include lodgepole pine, redwood, western larch, western juniper, and numerous other pine species, including knobcone, bishop, monterey, foxtail, limber, whitebark, and western white.

8. Major species of hardwoods are tanoak (CA and OR), red alder (OR and WA), bigleaf maple (OR and WA), black oak (CA), laurel (CA), canyon live oak (CA), Pacific madrone (CA), white oak (OR), and cottonwood (WA).

9. Partial-cut volume is estimated by comparing the measured volumes in 10-year intervals for the remeasured plots that have a record of partial-cut treatment in the most recent survey. For the other plots, we assigned the percentage of partial-cut portion of total volume according to the available information such as the treatment code that distinguishes “less than 20% removed” or “more than 20% removed” as well as county average percentage across the remeasured partially cut plots.

10. We use radial 10-year increment data on tree growth for each FIA plot to construct an approximation of plot-specific tree growth— $Δ {vol}_{nk (s_{j}) t}$ —which when multiplied by $P_{{ns}_{j} t}$ provides us with a direct measure of the marginal benefit of waiting to cut (see app. D for more details).

11. The inclusive value for the “no-cut” nest implicitly accounts for the risk of disturbance on the harvest decision, although this inclusive value is harder to interpret than in the clear-cut and partial-cut nests since the outcomes from choosing not to cut (natural disturbance or not) are not a direct choice by the landowner. Therefore, the inclusive values from the disturbance nest may not be justified as the utility gained by optimally choosing the best alternative, nor satisfy the consistency condition for utility maximization (Herriges and Kling 1996). See app. I for robustness checks for specifying the no-cut nest.

12. An alternative approach to estimation would be to set up a dynamic discrete choice model (e.g., Provencher 1995). However, while we can set up the harvest nest as a panel of repeated decisions, we only observe replanting decisions when the landowner has chosen to harvest (once for those that harvest between 2001 and 2014, never for those that do not harvest between 2001 and 2014). Thus, our data do not include temporal changes in the key explanatory variables of interest (climate) or panel data in the key behavioral variable of interest (replanting) that would make dynamic discrete choice estimation more informative than our current approach (Aguirregabiria and Mira 2010).

13. FIA surveys a fraction of all plots (10%) every year. Therefore, remeasurements occur every 10 years. For those remeasured, dependent variables reflect data from both measurements over a 10-year time frame.

14. Remote sensing data distinguishes neither different tree species nor detailed site characteristics.

15. Viability score values near zero indicate a low suitability while those near 1.0 indicate a suitability so high that the species is nearly always present in that climate. Although the score below 0.5 indicates little chance of survival, we use the score of 0.3 as a cut-off point whether the species is included in a choice set, to account for the error disclosed in Crookston et al. (2010).

16. Growing season months are those that have growing degrees days above 10°C (50°F), which are determined at a regional level that represents varying climate zones. Regional climate data are from National Oceanic and Atmospheric Administration (NOAA) National Climatic Data Center.

17. The downscaled data at 1 km resolution was obtained from the ClimateWNA model developed by the Center for Forest Conservation Genetics at the University of British Columbia (Wang et al. 2012), which is available at https://adaptwest.databasin.org/pages/adaptwest-climatewna.

18. We thank Crookston for providing us with plot-level viability scores for our study region. Appendix H shows maps of viable plots for Douglas-fir and hardwoods across the Pacific states.

19. The CCSM4 climate model is ranked as one of the best of 41 global climate models (GCM) as to the credibility of predicting the future climate according to the models’ abilities to reproduce the observed metrics (Rupp et al. 2013). We also simulated the landscape using an alternative climate projection of an ensemble of 15 GCMs instead of using one GCM, but the results show little changes from our original results.

20. Some have argued that carbon rents in forestry should account for the issue of additionality: pay only for the amount of carbon that would not have been sequestered without the carbon pricing policy. Additionality payments have a basic problem stemming from asymmetric information, which is that only the seller knows whether she would have undertaken the activity in the absence of a payment for the offset (Mason and Plantinga 2013). Implementing additionality payments in forest management presents challenging trade-offs between efficiency and budget concerns; recent work has proposed the use of a general land tax (Tahvonen and Rautiainen 2017) but has not considered how to implement the mechanism in a model where landowners can alter their replanting choice. Incorporating additionality into our model of replanting choices is beyond the present scope of our analysis.

21. Sohngen and Tian project that forest prices rise more slowly with climate change, largely because climate change is expected to increase overall global forest productivity.

22. The static expectations assumption is relaxed somewhat by our inclusion of mean temperature forecasts in the replanting nests and the simulation. However, as discussed in sec. 4, our econometric estimates find only limited evidence that current replanting decisions respond to future mean temperature forecasts.

23. This is consistent with current advice from Oregon State University Extension Service, which recommends that landowners should consider planting ponderosa pine or hardwoods where Douglas-fir mortality is occurring in western Oregon (Pokorny 2018).

24. Examples are Vaux’s swift, rufous hummingbird, sooty grouse, fisher, white-headed woodpecker, black-headed woodpecker, Sierra Nevada red fox, wolverine, lynx, Yuma myotis, and long-legged myotis.

25. Example hardwoods-only specialists are yellow-billed magpie; example hardwoods-ponderosa pine only specialists are ringtail, pallid bat, and oak titmouse.

References

Adams, Darius M., Ralph J. Alig, Bruce A. McCarl, J. M. Callaway, and Steven M. Winnett. 1996. An analysis of the impacts of public timber harvest policies on private forest management in the United States. Forest Science 42 (3): 343–58.
First citation in article
Google Scholar
Aguirregabiria, Victor, and Pedro Mira. 2010. Dynamic discrete choice structural models: A survey. Journal of Econometrics 156 (1): 38–67.
First citation in article
Crossref Google Scholar
Auffhammer, Maximilian, Patrick Baylis, and Catherine H. Hausman. 2017. Climate change is projected to have severe impacts on the frequency and intensity of peak electricity demand across the United States. Proceedings of the National Academy of Sciences 114 (8): 1886–91.
First citation in article
Crossref Google Scholar
Bateman, Ian, Matthew Agarwala, Amy Binner, Emma Coombes, Brett Day, Silvia Ferrini, Carlo Fezzi, Michael Hutchins, Andrew Lovett, and Paulette Posen. 2016. Spatially explicit integrated modeling and economic valuation of climate driven land use change and its indirect effects. Journal of Environmental Management 181:172–84.
First citation in article
Google Scholar
Carbone, Jared C., and V. Kerry Smith. 2013. Valuing nature in a general equilibrium. Journal of Environmental Economics and Management 66 (1): 72–89.
First citation in article
Crossref Google Scholar
Carleton, T. A., and S. M. Hsiang. 2016. Social and economic impacts of climate. Science 353 (6304): aad9837-1–aad9837-15.
First citation in article
Google Scholar
Coops, Nicholas C., Robbie A. Hember, and Richard H. Waring. 2010. Assessing the impact of current and projected climates on Douglas-fir productivity in British Columbia, Canada, using a process-based model (3-PG). Canadian Journal of Forest Research 40 (3): 511–24.
First citation in article
Crossref Google Scholar
Coops, Nicholas C., and Richard H. Waring. 2011. Estimating the vulnerability of fifteen tree species under changing climate in northwest North America. Ecological Modelling 222 (13): 2119–29.
First citation in article
Crossref Google Scholar
Crookston, Nicholas L., Gerald E. Rehfeldt, Gary E. Dixon, and Aaron R. Weiskittel. 2010. Addressing climate change in the forest vegetation simulator to assess impacts on landscape forest dynamics. Forest Ecology and Management 260 (7): 1198–1211.
First citation in article
Crossref Google Scholar
Dennis, Donald F. 1990. A probit analysis of the harvest decision using pooled time-series and cross-sectional data. Journal of Environmental Economics and Management 18 (2, pt. 1): 176–87.
First citation in article
Google Scholar
Deschênes, Olivier, and Michael Greenstone. 2007. The economic impacts of climate change: Evidence from agricultural output and random fluctuations in weather. American Economic Review 97 (1): 354–85.
First citation in article
Crossref Google Scholar
Ekholm, T. 2016. Optimal forest rotation age under efficient climate change mitigation. Forest Policy and Economics 62:62–68.
First citation in article
Crossref Google Scholar
Fankhauser, Sam. 2017. Adaptation to climate change. Annual Review of Resource Economics 9 (1): 209–30.
First citation in article
Crossref Google Scholar
Fezzi, Carlo, Amii R. Harwood, Andrew A. Lovett, and Ian J. Bateman. 2015. The environmental impact of climate change adaptation on land use and water quality. Nature Climate Change 5:255–60.
First citation in article
Crossref Google Scholar
Fulton, S., and B. West. 2002. Forestry impacts on water quality. In Southern forest resource assessment, ed. D. Wear and John G. Greis, chap. 21. General technical report SRS-53. Asheville, NC: US Department of Agriculture, Forest Service, Southern Research Station: 628.
First citation in article
Google Scholar
Graff Zivin, Joshua, and Matthew Neidell. 2014. Temperature and the allocation of time: Implications for climate change. Journal of Labor Economics 32 (1): 1–26.
First citation in article
Link Google Scholar
Grotta, Amy T., Janean H. Creighton, Christopher Schnepf, and Sylvia Kantor. 2013. Family forest owners and climate change: Understanding, attitudes, and educational needs. Journal of Forestry 111 (2): 87–93.
First citation in article
Crossref Google Scholar
Guo, Christopher, and Christopher Costello. 2013. The value of adaption: Climate change and timberland management. Journal of Environmental Economics and Management 65 (3): 452–68.
First citation in article
Crossref Google Scholar
Hanewinkel, Marc, Dominik A. Cullmann, Mart-Jan Schelhaas, Gert-Jan Nabuurs, and Niklaus E. Zimmermann. 2012. Climate change may cause severe loss in the economic value of European forest land. Nature Climate Change 3 (3): 203–7.
First citation in article
Google Scholar
Hannah, L., C. Costello, C. Guo, L. Ries, C. Kolstad, D. Panitz, and N. Snider. 2011. The impact of climate change on California timberlands. Climatic Change 109 (S1): 429–43.
First citation in article
Crossref Google Scholar
Hartman, R. S. 1988. Self-selection bias in the evolution of voluntary energy conservation programs. Review of Economics and Statistics 70:448.
First citation in article
Crossref Google Scholar
Herriges, J. A., and C. L. Kling. 1996. Testing the consistency of nested logit models with utility maximization. Economics Letters 50:33–39.
First citation in article
Crossref Google Scholar
Im, Eun Ho, Darius M. Adams, and Gregory S. Latta. 2007. Potential impacts of carbon taxes on carbon flux in western Oregon private forests. Forest Policy and Economics 9 (8): 1006–17.
First citation in article
Crossref Google Scholar
Iverson, Louis R., and Donald McKenzie. 2013. Tree-species range shifts in a changing climate: Detecting, modeling, assisting. Landscape Ecology 28 (5): 879–89.
First citation in article
Crossref Google Scholar
Lawler, J. J., D. J. Lewis, E. Nelson, A. J. Plantinga, S. Polasky, J. C. Withey, D. P. Helmers, S. Martinuzzi, D. Pennington, and V. C. Radeloff. 2014. Projected land-use change impacts on ecosystem services in the United States. Proceedings of the National Academy of Sciences 111 (20): 7492–97.
First citation in article
Crossref Google Scholar
Lee, Dug Man, and Kenneth S. Lyon. 2004. A dynamic analysis of the global timber market under global warming: An integrated modeling approach. Southern Economic Journal 70 (3): 467–89.
First citation in article
Crossref Google Scholar
Lewis, David J., and Andrew J. Plantinga. 2007. Policies for habitat fragmentation: Combining econometrics with GIS-based landscape simulations. Land Economics 83 (2): 109–27.
First citation in article
Crossref Google Scholar
Lubowski, Ruben N., Andrew J. Plantinga, and Robert N. Stavins. 2006. Land-use change and carbon sinks: Econometric estimation of the carbon sequestration supply function. Journal of Environmental Economics and Management 51 (2): 135–52.
First citation in article
Crossref Google Scholar
Lyon, Kenneth S., and Roger A. Sedjo. 1983. An optimal control theory model to estimate the regional long-term supply of timber. Forest Science 29 (4): 798–812.
First citation in article
Google Scholar
Mason, Charles F., and Andrew J. Plantinga. 2013. The additionality problem with offsets: Optimal contracts for carbon sequestration in forests. Journal of Environmental Economics and Management 66 (1): 1–14.
First citation in article
Crossref Google Scholar
Massetti, Emanuele, and Robert Mendelsohn. 2018. Measuring climate adaptation: Methods and evidence. Review of Environmental Economics and Policy 12 (2): 324–41.
First citation in article
Link Google Scholar
Mathys, Amanda S., Nicholas C. Coops, and Richard H. Waring. 2017. An ecoregion assessment of projected tree species vulnerabilities in western North America through the 21st century. Global Change Biology 23 (2): 920–32.
First citation in article
Crossref Google Scholar
McKenney, D. W., J. H. Pedlar, J. Yang, A. Weersink, and G. Lawrence. 2015. An economic analysis of seed source options under a changing climate for black spruce and white pine in Ontario, Canada. Canadian Journal of Forest Research 45:1248–57.
First citation in article
Crossref Google Scholar
Mendelsohn, Robert, William D. Nordhaus, and Daigee Shaw. 1994. The impact of global warming on agriculture: A Ricardian analysis. American Economic Review 84 (4): 753–71.
First citation in article
Google Scholar
Mendelsohn, R., I. C. Prentice, O. Schmitz, B. Stocker, R. Buchkowski, and B. Dawson. 2016. The ecosystem impacts of severe warming. American Economic Review 106 (5): 612–14.
First citation in article
Crossref Google Scholar
Nelson, Erik, Stephen Polasky, David J. Lewis, Andrew J. Plantinga, Eric Lonsdorf, Denis White, David Bael, and Joshua J. Lawler. 2008. Efficiency of incentives to jointly increase carbon sequestration and species conservation on a landscape. Proceedings of the National Academy of Sciences 105 (28): 9471–76.
First citation in article
Crossref Google Scholar
Pokorny, Kym |For The Oregonian/OregonLive. 2018. Oregon conifers hit hard by years of drought; here’s how to keep trees healthy. https://www.oregonlive.com/hg/index.ssf/2018/09/oregon_conifers_hit_hard_by_ye.html.
First citation in article
Google Scholar
Polyakov, Maksym, David N. Wear, and Robert N. Huggett. 2010. Harvest choice and timber supply models for forest forecasting. Forest Science 56 (4): 344–55.
First citation in article
Google Scholar
Prasad, Anantha M., Judith D. Gardiner, Louis R. Iverson, Stephen N. Matthews, and Matthew Peters. 2013. Exploring tree species colonization potentials using a spatially explicit simulation model: Implications for four oaks under climate change. Global Change Biology 19 (7): 2196–2208.
First citation in article
Crossref Google Scholar
Prestemon, Jeffrey P., and David N. Wear. 2000. Linking harvest choices to timber supply. Forest Science 46 (3): 377–89.
First citation in article
Google Scholar
Provencher, Bill. 1995. Structural estimation of the stochastic dynamic decision problems of resource users: An application to the timber harvest decision. Journal of Environmental Economics and Management 29 (3): 321–38.
First citation in article
Crossref Google Scholar
———. 1997. Structural versus reduced-form estimation of optimal stopping problems. American Journal of Agricultural Economics 79 (2): 357–68.
First citation in article
Crossref Google Scholar
Reed, William J. 1984. The effects of the risk of fire on the optimal rotation of forest. Journal of Environmental Economics and Management 11:180–90.
First citation in article
Crossref Google Scholar
Rehfeldt, Gerald E., Barry C. Jaquish, Cuauhtémoc Sáenz-Romero, Dennis G. Joyce, Laura P. Leites, J. Bradley St. Clair, and Javier López-Upton. 2014. Comparative genetic responses to climate in the varieties of Pinus ponderosa and Pseudotsuga menziesii: Reforestation. Forest Ecology and Management 324 (July): 147–57.
First citation in article
Crossref Google Scholar
Rupp, David E., John T. Abatzoglou, Katherine C. Hegewisch, and Philip W. Mote. 2013. Evaluation of CMIP5 20th century climate simulations for the Pacific Northwest USA. Journal of Geophysical Research: Atmospheres 118 (19): 10884–906.
First citation in article
Crossref Google Scholar
Sanford, Todd, Peter C. Frumhoff, Amy Luers, and Jay Gulledge. 2014. The climate policy narrative for a dangerously warming world. Nature Climate Change 4 (3): 164–66.
First citation in article
Crossref Google Scholar
Schlenker, Wolfram, W. Michael Hanemann, and Anthony C. Fisher. 2006. The impact of global warming on US agriculture: An econometric analysis of optimal growing conditions. Review of Economics and Statistics 88 (1): 113–25.
First citation in article
Google Scholar
Seo, S. N., and R. Mendelsohn. 2008. An analysis of crop choice: Adapting to climate change in South American farms. Ecological Economics 67:109–16.
First citation in article
Crossref Google Scholar
Severen, Christopher, Christopher Costello, and Olivier Deschênes. 2018. A forward-looking Ricardian approach: Do land markets capitalize climate change forecasts? Journal of Environmental Economics and Management 89:235–54.
First citation in article
Crossref Google Scholar
Sohngen, Brent, and Robert Mendelsohn. 1998. Valuing the impact of large-scale ecological change in a market: The effect of climate change on U.S. timber. American Economic Review 88 (4): 686–710.
First citation in article
Google Scholar
Sohngen, Brent, and Xiaohui Tian. 2016. Global climate change impacts on forests and markets. Forest Policy and Economics 72:18–26.
First citation in article
Crossref Google Scholar
Susaeta, A., S. J. Chang, D. R. Carter, and P. Lal. 2014. Economics of carbon sequestration under fluctuating economic environment, forest management and technological changes: An application to forest stands in the southern United States. Journal of Forest Economics 20:47–64.
First citation in article
Crossref Google Scholar
Stavins, Robert N. 1999. The costs of carbon sequestration: A revealed-preference approach. American Economic Review 89 (4): 994–1009.
First citation in article
Crossref Google Scholar
Tahvonen, Olli, and Aapo Rautiainen. 2017. Economics of forest carbon storage and the additionality principle. Resource and Energy Economics 50:124–34.
First citation in article
Crossref Google Scholar
Train, Kenneth E. 2009. Discrete choice methods with simulation. Cambridge: Cambridge University Press.
First citation in article
Google Scholar
van Kooten, G. C., C. S. Binkley, and G. Delcourt. 1995. Effect of carbon taxes and subsidies on optimal forest rotation age and supply of carbon services. American Journal of Agricultural Economics 77:365.
First citation in article
Crossref Google Scholar
Wang, T., A. Hamann, D. L. Spittlehouse, and T. Q. Murdock. 2012. ClimateWNA: High-resolution spatial climate data for western North America. Journal of Applied Meteorology and Climatology 51:16–29.
First citation in article
Crossref Google Scholar
Wilcove, D. S., D. Rothstein, J. Dubow, A. Phillips, and E. Losos. 1998. Quantifying threats to imperiled species in the United States. BioScience 48:607–15.
First citation in article
Crossref Google Scholar