Potential impacts of insect-induced harvests in the mixed forests of New England

Forest insects and pathogens (FIPs) have significant impacts on U.S. forests, each year affecting an area nearly three times the area of wildfires and timber harvesting combined. We surveyed family forest owners (FFOs) in the northeastern U.S. and 84% of respondents indicated they would harvest in at least one of the presented FIP infestation scenarios. This harvest response represents a potentially significant shift in the timing, extent, and species selection of harvesting in the Northeast. Here we used the landowner survey, regional forest inventory data, and characteristics of the emerald ash borer (EAB, Agrilus planipennis) invasion to examine the potential for a rapidly spreading FIP to alter harvest regimes and affect regional forest conditions. Twenty-five percent of the FFO parcels in the Connecticut River Watershed in New England are likely to be harvested in response to EAB within 10 years. This prediction represents an increase in harvest frequencies, from 2.9%/yr (historically) to 3.7%/yr, on FFO woodlands. At typical harvest intensities, this would result in 13% of the total aboveground biomass removed through these harvests, with 81% of that biomass from species other than ash, creating a forest disturbance that is over twice the magnitude of the disturbance from EAB alone.

Timber harvesting is the primary driver of forest disturbance in the northeast (Brown et al. 41 2018); therefore, this shift in the harvest regime would change regional patterns of disturbance 42 with potentially profound impacts. Here we use the survey and regional forest inventory data to 43 examine the potential for an emerging and rapidly spreading FIP, the emerald ash borer (Agrilus all harvests, salvage harvests can alter forest development trajectories and change structural 48 legacies with lasting impacts on biodiversity and ecosystem services (Leverkus et al. 2018). 49 Salvage harvests usually occur after the disturbance, but occasionally, and particularly in the 50 case of FIPs, occur preemptively in an attempt to mitigate future damage or value loss, or slow 51 FIP spread (Waring and O'Hara 2005). Here we use the term "FIP harvest" to refer to any 52 harvest that was initiated either in anticipation of or in response to damage from a FIP. One 53 concern with FIP harvesting in mixed hardwood forests, like those in the Northeast, is the 54 phenomenon of harvest "by-catch", or the non-host tree species harvested with the FIP-host 55 species to enhance the commercial viability of the harvest or achieve another silvicultural 56 objective (e.g. regeneration of a desirable mix of species). Irland et al. (1988) reported 57 landowners harvesting to mitigate damages from spruce budworm (Choristoneura fumiferana) 58 outbreaks in Maine, and those harvests included tree species beyond the budworm's primary 59 regional hosts. Similarly, Kizlinski et al. (2002)  FIPs and subsequent harvest, as well as determine which tree species co-occur with ash and 120 therefore may be most affected by the harvest response to EAB (Appendix A). For this study, 121 we only used the distributions of white ash (Fraxinus americana) and black ash (Fraxinus 122 nigra), and while these are the two most abundant ash species in the CTRW, white ash makes up 123 99.96% of the total ash biomass (Table 1). 124 Determining harvest response to EAB 125 In this region, a majority of FFO harvests are low intensity (e.g. non-commercial) harvests, but 126 are ecologically important drivers of forest change (Thompson et al. 2017, Belair and Ducey 127 2018). Therefore, for modeling the harvest response to EAB, we assumed the range of possible 128 harvest responses included commercial and non-commercial harvests. We also assumed that 129 EAB was present throughout the study area (at its current rate of spread, this assumption will 130 likely be met within one decade). Following Holt et al. would harvest. For this EAB application, the relative abundance of ash on the parcel was used as 141 the mortality percent and we made the conservative assumption of 8 years to mortality from 142 initial infestation for those ash trees. The probabilities of harvest were then stochastically used 143 to predict which parcel owners might harvest in response to EAB. The process of assigning 144 AFTs, probabilities of harvest, and finally harvest locations, was replicated 10 times to capture 145 the variability in assigning harvest locations. 146

Quantifying cumulative impacts of EAB 147
We used the forest composition map and the 10 simulated harvest maps to estimate species AGB 148 removal in EAB-initiated FFO harvests. To account for the uncertainty regarding the specific 149 silvicultural choices made by foresters and loggers, we examined three alternative harvest 150 scenarios and report the total AGB and species-specific AGB that could be harvested in response 151 to EAB. Our estimates include the AGB of merchantable timber trees (i.e., larger than 23 cm (9 152 in) diameter at breast height (dbh) for softwoods and 28 cm (11 in) dbh for hardwoods) and only 153 parcels that met the criteria for our survey (i.e., FFO parcels > 4.05 ha). We also set a minimum 154 forest area size of 3.15 ha on mixed land use parcels to ensure enough forest for measurable tree 155 removal. We included three harvest scenarios, described below. 156 Scenario A: Typical harvest intensities and no species preferences 157 Scenario A assumed that the presence of a FIP is an exogenous force that initiates a typical FFO 158 harvest, rather than a reactionary salvage harvest. As some state agencies and interest groups 159 recommend removing healthy ash just prior to infestation (e.g., see NHBugs.org), some 8 landowners are choosing to pre-emptively harvest ash, which may indeed result in a more 161 planned and typical harvest than a reactionary harvest after infestation. Furthermore, the 162 harvests in this scenario did not differentially affect tree species based on their type or value; 163 rather, the proportion of AGB removed, or harvest intensity, was applied equally to all species 164 with trees of merchantable timber size on the parcel. To determine what constitutes a "typical 165 harvest" intensity for FFO harvests in the CTRW, we used the harvest intensity distributions 166 from re-measured FIA plots on FFO lands in the CTRW states that experienced tree removal 167 between measurements. We included all types of removals, from very low intensity removals to 168 clear-cuts, to define these distributions (Appendix B). Harvest intensities for individual parcels 169 were then drawn from these distributions. All merchantable ash was first removed, then any 170 remaining timber was removed proportionally from all remaining species until the target harvest 171 intensity was met. Where the total volume of ash on a parcel exceeded the necessary volume for 172 the target harvest intensity, all merchantable ash was removed, but no more. 173

Scenario B: Typical harvest intensities and species preferences 174
Scenario B differentially harvests tree species based on patterns observed in the historical data. 175 In this scenario, the harvest intensities were the same as in Scenario A, but the amount removed 176 for each species (other than ash) was weighted to match the observed harvest distributions from 177 the re-measured FIA plots. Weights were calculated by dividing the proportion of each species 178 removed from the initial measurement to the re-measurement of a plot by the overall harvest 179 intensity for that plot and then averaged over all harvested plots (Appendix C). 180 Scenario C: Decreased harvest intensity and increased species preference 181 In Scenario C, we examined the possibility that harvest prompted by EAB will be less intense 182 than typical harvests, supposing that many of the harvests were reactionary and not planned with 183 harvest volumes as an objective. Therefore, the harvest intensities for each parcel were reduced 184 by 50%. In addition, the weighting of species removal was intensified to approximate a behavior 185 where preferable trees are selectively chosen in a low-intensity FIP harvest so the landowner can 186 recover the operational cost of harvesting while removing the fewest number of trees beyond the 187 affected ash. To reweight the species, weights of preferential trees (weight > 1) were increased 188 by 10% and the weights of non-preferential trees (weight < 1) were decreased by 10%. 189

RESULTS 190
Determining harvest response to EAB 191 In total, 80,010 FFO parcels met the area requirements for harvest consideration, and of these 192 average harvest intensity applied to harvest parcels in the northern region was 45%±28%, and 202 the average harvest intensity applied to the southern region was 27%±22%. 203 Quantifying cumulative impacts of EAB 204 We found 8.2 Tg of merchantable timber size ash biomass on harvestable FFO parcels in the 205 CTRW; up to 99% of which is expected to be killed by EAB (Klooster et al. 2018). 206 Approximately 54% (4.4 Tg) of that ash is expected to be removed from the landscape through 207 harvest in all scenarios (Table 1). In Scenario A, 18.8 Tg of AGB was harvested from co-208 occurring species (or as by-catch) in addition to the ash (Table 1). Including the ash removal, the 209 total AGB harvested in response to EAB equates to 13% of all AGB in the harvestable FFO 210 parcels in the CTRW (with or without ash). In Scenario B, a similar amount of biomass was 211 removed as in Scenario A; however, the amount of each species removed differed. In this 212 scenario, yellow birch (Betula alleghaniensis), paper birch (B. papyrifera), and black cherry 213 (Prunus serotina) incur the largest potential losses, up to 17% of their AGB, due to preferential 214 harvesting and co-occurrence with ash (Table 1 and Appendix A). In Scenario C, with decreased 215 harvest intensity only 9.0 Tg of biomass was harvested from co-occurring species. However, a 216 few species, such as yellow birch, were removed at higher than average proportions. 217

DISCUSSION 218
We used forest owner survey data and a landowner typology to examine three scenarios 219 describing the harvest response to EAB; they suggest that harvest frequency could increase 28% 220 above the recent trends in harvesting (sensu Thompson et al. 2017). Twenty-five percent of 221 parcels, representing 37% of the FFO forested area, within the CTRW were predicted to be 222 harvested in response to EAB. Given the spread rate of EAB and its current locations within the 223 watershed, most of this harvest is predicted to take place in the next decade. Assuming this 224 future, the annual probability of harvest in response to EAB could be 3.7% for FFO lands in the 225 CTRW for the next 10 years. Thompson et al. (2017) found that in the 20-state northeastern 226 region of the FIA, the annual probability of harvest on non-corporate private woodlands was harvest response in these scenarios created a disturbance, in terms of total AGB affected, more 233 than twice the magnitude of the expected disturbance of EAB and included a larger number of 234 species. In scenario C, where harvest intensities are lower, 65% of the removed biomass was 235 from species other than ash, creating a disturbance 1.5 times the magnitude of EAB alone. Both 236 co-occurrence with ash and silvicultural prescription influenced the scope of the disturbance for 237 individual species. For example, 81% of yellow birch (B. alleghaniensis) is co-located with ash. 238 Since yellow birch is also a preferred species for harvest, 17% of the merchantable timber-size 239 yellow birch trees were harvested in the typical harvest intensities and species preference 240 scenario (B). If average harvest intensities are lower for the FIP harvests but target preferential 241 species (e.g., Scenario C), the cumulative impacts of EAB is less but preferred merchantable 242 species that co-occur with ash (e.g. yellow birch) are removed proportionally more than species 243 of less value or that do not co-occur with ash as frequently (e.g. black birch, B. lenta). 244 In all scenarios, more frequent harvests have far-reaching ecological impacts. Harvesting in 245 response to a FIP with a specific host(s) alters the species composition of the landscape by 246 differentially impacting species that co-occur with the host. The by-catch in FIP harvesting will 247 be especially prevalent in the mixed forests of the Northeast when harvests are planned with 248 timber volumes and/or larger regeneration openings as objectives, since, one host species often 249 will not be enough to meet target volumes or silvicultural objectives.

12
The survey we used to develop and apply the landowner typology did not ask questions about the 252 type or intensity of harvesting that a FIP might induce. To account for this uncertainty, we 253 presented three scenarios of varying harvest intensity and species removal. Further investigation 254 into the intensity and types of silvicultural prescriptions applied in the harvest response to EAB 255 is necessary to quantify these effects, particularly from the viewpoint of the foresters and loggers 256 who are completing the harvests. Additionally, it is unknown if the harvest response may be 257 limited by the capacity of foresters, loggers, and markets. We also expect that there is some 258 interplay between harvest unrelated to EAB, FIP harvest in response to EAB, forest growth 259 dynamics, and climate change. To assess specifically how FIP harvests may accelerate and 260 intensify forest disturbance and change forest growth and composition over time, future research 261 should explore these dynamics with a spatially and temporally explicit representation of 262 background harvest and FIP harvest initiated by EAB.

APPENDICES Appendix A: Co-occurrence of species with ash
The forest composition map was used to assess which species co-occur with ash on all FFO parcels. Table A1. Tree species that most commonly co-occur on parcels with ash (Fraxinus spp.). Total merchantable timber is the total amount of species-specific aboveground biomass (AGB) in the CTRW. Average co-occurrence is 74%. The percent of total AGB that co-occurs with ash represents the amount of that species that occurs on parcels with ash; e.g., 84% of all black cherry in the CTRW occurs on parcels with ash.   Figure 2). For this specific application using EAB, the mortality percent was the relative abundance of ash for each parcel and we chose a conservative time to death of 8 years for ash at the time of initial infestation. The distribution of harvest probabilities was unique for each landowner type class ( Figure B2). The probability of harvest was then used to stochastically assign harvest or no harvest to each parcel ( Figure 1).

Species
Subsequently, to mimic typical harvest intensities across the landscape, we assigned a harvest intensity (AGB removed/AGB prior to harvest) to each harvest parcel. We used remeasured FIA    Weights were based on species removals from individual FIA plots in the states of the CTRW.
Weights were calculated by dividing the proportion of each species removed from the initial measurement to the re-measurement of each FIA plot by the harvest intensity of that plot. These proportions were then averaged over all harvested plots. Species removed in higher proportions than the average received a weight > 1, while species removed in lower proportions received a weight < 1. For example, if sugar maple (Acer saccharum) existed on a harvested plot, less of it was harvested than yellow birch (Betula alleghaniensis).