Long-Term Impacts of Invasive Herbivores on Tree Physiology, Growth, and Phenology: A Whole-Tree Perspective

Insect herbivores play an essential role in structuring plant communities and species interaction therein. Plant response to herbivory, particularly to non-native insects, can be difficult to predict. A diverse array of feeding strategies, including leafchewing, wood-boring, and piercing-sucking, leads to varied plant responses following attack. Piercing-sucking insects are known to alter source-sink dynamics but are relatively understudied, particularly in woody plants. Two piercing-sucking invasive insects, hemlock woolly adelgid (Adelges tsugae; ‘HWA’) and the elongate hemlock scale (Fiorinia externa; ‘EHS’), are commonly found on eastern hemlock (Tsuga canadensis; ‘hemlock’) in the eastern United States. Hemlock, a native shadetolerant conifer, provides unique habitat for a range of biota and plays an important role in structuring ecosystems, but is threatened throughout much of its range because of HWA. HWA drives rapid decline in tree health, whereas EHS rarely kills trees. The individual and interactive impacts of HWA and EHS on resource allocation, phenology, and metabolite profiles were explored following two and four years of infestations. HWA-infested trees, regardless of EHS presence, had relatively more biomass belowground and less aboveground biomass. Consistent needle desiccation and drop indicative of HWA infestation explains these allocation changes. EHS did not drive changes in biomass allocation. HWA-infested trees broke bud on average three days later than HWA-free trees and new flush production (grams/day) in early spring was 30% less compared to HWA-free trees. Although EHS and HWA both impacted primary metabolites, the effects of HWA are more pronounced. While EHS has virtually no impact, HWA substantially alters resource acquisition and allocation in eastern hemlock. Assessing whole-plant impacts of two invasive piercing-sucking insects on a native woody plant following long-term experimental infestations in a long-lived conifer provides a unique contribution to the literature. The lack of interaction between HWA and EHS at a whole-plant level, which conflicts with prior branch-level studies, reinforces the importance of considering long-term impacts in an ecologically relevant setting. The accompanying appendices contain additional details about canopy closure (Appendix S1) and statistical models (Appendix S2).

and drop indicative of HWA infestation explains these allocation changes. EHS did not drive changes in biomass allocation. HWA-infested trees broke bud on average three days later than HWA-free trees and new flush production (grams/day) in early spring was 30% less compared to HWA-free trees. Although EHS and HWA both impacted primary metabolites, the effects of HWA are more pronounced. While EHS has virtually no impact, HWA substantially alters resource acquisition and allocation

INTRODUCTION
The ubiquity and relative immobility of autotrophs in terrestrial and aquatic communities makes them a tempting target for a diverse array of herbivorous insects (Karban andBaldwin 1997, Stam et al. 2014). The impact of these attackers on the morphology, physiology, and growth of individual plants can substantially alter community structure (Crawley 1989, Marquis 2004. Researchers have made significant strides in understanding how insects individually (Kessler and Baldwin 2002) and jointly (Stam et al. 2014) affect their host plants. As the number of herbivore invasions rise and their cumulative ecosystem impacts increase, it is increasingly important to understand the individual and combined effect of invasive herbivores on plant fitness (Lovett et al. 2006, Gandhi andHerms 2010).
Because herbivores differ in their phenology, their attacks on the same plant can be simultaneous or sequential. Across different spatial and temporal scales, the resulting damage can be the sum of individual herbivore effects (e.g. Knochel et al 2010); it is also possible, however, for the single-species effects to interact in a non-additive manner (Kaplan and Denno 2007). Non-additive impacts on a shared host plant are particularly likely when early-arriving herbivores induce changes in host plant physiology and chemistry (Fournier et al. 2006, Morris et al. 2007, Pieterse and Dicke 2007, Stam et al. 2014). An early-season herbivore can increase damage by a late-arriving herbivore if early-season herbivory makes the plant more susceptible to pathogens (Wallin and Raffa 2001) or attenuates plant defenses (Soler et al. 2012).
Alternately, an early-arriving herbivore may reduce the damage caused by a laterarriving herbivore through the induction of plant defense or changes in plant quality (Hunter 1987). Understanding the nuances of such herbivore interactions is especially critical in cases involving invasive species with community-or ecosystem-level impacts. The presence of several such species could lead to invasional meltdown (Simberloff 1999), for instance, or generate invasional interference (Yang et al. 2011, Rauschert andShea 2012).
Sap-feeding herbivores are a diverse group of insects whose impact on woody plants can equal or exceed that of defoliators (Zvereva et al. 2010). Woody plants are less likely than herbaceous plants to respond to an attack with compensatory growth, instead show a reduction of photosynthate production (Schowalter 1981).
Additionally, woody plants are less tolerant of sap-feeders than defoliators, yet research is skewed towards examining defoliator impacts on plant health (Zvereva et al. 2010). Studies addressing the impacts of multiple sap-feeding herbivores often utilize naturally-infested plants (e.g., Dungan et al. 2007, Grégoire et al. 2015, Karban 1980, an approach that cannot be used to experimentally assess non-additive effects (Nykänen and Koricheva 2004). As a result, few studies have examined how simultaneous and sequential attacks by sap-feeding insects impact woody plants, especially over the multi-year timescales most appropriate to assessments of their health (Zvereva et al. 2010).
While numerous stand-alone metrics can be used to quantify plant health, the complementary and interactive nature of many plant responses highlights the appeal of a 'whole-plant' approach to herbivory studies. This approach simultaneously measures herbivore-induced alterations in factors such as growth, metabolism, and resource allocation; while logistically complex, such work is essential to identifying synergistic or compensatory responses across multiple 'compartments' (e.g., foliar versus woody biomass, or above-versus belowground tissue). Work by Moreira et al. (2015) illustrated the complex plant responses induced by two folivores feeding sequentially on lima bean (Phaseolus lunatus): the order of herbivore arrival influenced some metrics of plant reproduction (seed mass, germination) but not others (seed number).
A whole-plant approach also makes it easier to assess how herbivore-induced change in growth or metabolism affect plant life history events, including biomass allocation and phenology. The latter is especially important for forest understory plants that rely on high photosynthetic rates in early spring before canopy leaf-out; herbivore-induced delays in bud break may put them at particular risk.
We report the results of a four-year experiment assessing the responses of a foundational tree species to the individual and combined presence of two invasive sapfeeding herbivores. We describe how chronic herbivory by the hemlock woolly adelgid (Adelges tsugae, 'HWA') and elongate hemlock scale (Fiorina externa, 'EHS') affected the growth, physiology, and phenology of eastern hemlock (Tsuga canadensis, 'hemlock') saplings. Eastern hemlock is a late successional foundational tree species that grows well in high-shade forest understories (Ellison et al. 2005. The fact that both herbivore species are sessile, along with previous research into this interaction , Gómez et al. 2012, Miller-Pierce and Preisser 2012, Domec et al. 2013) make this an ideal model system for exploring whole-plant impacts of chronic herbivory on a woody plant. Over the course of four years, hemlock saplings planted into a deciduous forest understory were individually, simultaneously, or sequentially inoculated with neither, one, or both herbivores. This design allows us to explore the multi-year impact of herbivore identity, single-versus multi-species infestations, and consumer priority effects on the growth, metabolism, and phenology of a long-lived woody plant. Our results illustrate the complex and multi-faceted impacts of chronic herbivory on a long-lived tree species, and highlight the importance of long-term experimental manipulations for exploring the interplay between herbivores and woody plants.

Natural History
Eastern hemlock (Tsuga canadensis, 'hemlock') is a late-successional, shadetolerant conifer that relies on early-and late-season carbon capture that occurs prior to spring hardwood leaf-out and following fall hardwood leaf-loss (Hadley and Schedlbauer 2002). It plays a critical role in structuring forests in the eastern United States, and has been identified as a 'foundational species' in these ecosystems (Ellison et al. 2005).
Hemlock woolly adelgid (Adelges tsugae, 'HWA') was introduced from Japan to the eastern United States in the 1950s (Havill et al. 2006); its invaded range now extends from Georgia northwards to Maine (Morin et al. 2009, Gómez et al. 2015. Mobile first-instar 'crawlers' settle at the base of hemlock needles and become sessile adults that extract photosynthate from xylem ray parenchyma cells (Young et al. 1995). It poses a serious threat to hemlock, with many attacked trees dying within 5-10 years of infestation (Orwig et al. 2002).

Elongate hemlock scale (Fiorinia externa; 'EHS') arrived in the eastern United
States from Japan in the early 1900s (Sasscer 1912) and co-occurs with HWA on hemlock throughout the eastern United States (Lambdin et al. 2005 Following planting, each tree was randomly assigned to a treatment (Table 1); a small number of trees were subsequently reassigned so that each row and column contained each treatment. Inter-plant dispersal of HWA and EHS is most likely prior to spring leaf-out, when both sub-canopy wind velocities and crawler densities are high (McClure 1989). Each spring, we simulated this yearly dispersal event by inoculating each tree with foliage infested with the appropriate insect; herbivore-free trees were 'inoculated' with uninfested foliage in order to control for handling.
Inoculations were conducted using a standard protocol (Butin et al. 2007); because HWA emerges earlier than EHS, inoculations were conducted in May and June, respectively.
Starting in 2011, trees in three treatments were annually inoculated with HWA only, EHS only, or both insects for four years (HWA-4, EHS-4, and Both-4, respectively). Starting in 2013, some HWA-only and some EHS-only trees were thereafter annually inoculated with both insects, creating two 'priority effect' treatments (HWAàBoth, EHSàBoth). In 2013 we also began annual inoculations of previously-uninfested trees with HWA-only, EHS-only, or both insects for two years (HWA-2, EHS-2, Both-2). A subset of trees remained herbivore-free throughout the experiment (Control).
Insect densities were assessed twice yearly, in early spring and late fall,  Soltis et al. 2015). As in prior work, the densities of both HWA and EHS were higher in single-species treatments than when they co-occurred (150% higher for HWA and 50% higher for EHS).
Between 2011-2015, we lost replicates to Hurricane Sandy, cross-treatment contamination, browsing by white tailed deer (Odocoileus virginianus), and isolated outbreaks of secondary pests (e.g., Oligonychus ununguis mites and Nucalaspis sp. scales). There were also a few trees in the single-herbivore treatments (i.e., treatments EHS-2, EHS-4, HWA-2, and HWA-4) whose persistently-low insect densities (<0.5 insects/cm; the bottom 15% of fall 2014 insect densities) may have obscured the impact of insect damage; we excluded these trees from our final harvest. The 88 remaining trees were marked for intensive monitoring in early spring prior to the May 2015 harvest.

Early spring monitoring
In early April 2015, we haphazardly selected and marked three branches per tree for assessments of insect density, bud phenology, and post-harvest biomass and chemical analyses. Each branch was marked and all insects present on it were counted (detailed below). This was completed prior to bud break and the emergence of HWA crawlers.
Between April 15-19th 2015, we measured gas exchange in one-year-old (2014 growth) foliage on the terminal end of each marked branch, for a total of three measurements per tree. All measurements were taken prior to bud break. We used a CIRAS-2 portable photosynthesis system (PP systems, Haverhill, MA, USA) with a 2.5 cm 2 cuvette and a CIRAS-2 LED light source of 1,500 µmol m -2 s -1 , a CO 2 concentration of 390 ppm, air flow rate at 350 cm 3 s -1 , and leaf temperature of 25°C.
All measurements occurred between 8:00 AM and 12:00 PM. After each measurement, the foliage inside the cuvette was photographed and ImageJ 1.44 (Abràmoff et al. 2004) was used to quantify needle area to account for blank space in the cuvette.
From April 30-May 16th, we monitored the three marked branches per tree for terminal bud break every other day. No trees broke bud prior to April 30th; any branches that had not broken bud by May 18th, the day that harvest began, were scored as broken on May 19th. We compared the timing of hemlock bud break to the bud break of co-occurring hardwoods using estimates of Normalized Vegetation Difference Index (NDVI) and Leaf Area Index (LAI) obtained from NASA MODIS (Moderate Resolution Imaging Spectroradiometer) (See Appendix S1 for details).

Harvest and biomass measurements
The 88 experimental trees were harvested between May 18-29 2015. Because of the time required for whole-tree excavation, we split the trees into 22 four-tree harvest groups, with each treatment represented in at least every third group.
Depending on the available person power, 1-3 groups were harvested daily. The growing window for each tree was estimated by subtracting the mean day of bud break for the three marked branches from its harvest date. This was used to calculate rate of new flush production.
Immediately prior to harvesting each tree, we recorded its height and trunk diameter five cm above the root ball. Following these measurements, the three previously-marked branches were each clipped at the base, placed in zip lock bags and stored on ice; to ensure we possessed sufficient plant material for chemical analyses, we also collected a fourth haphazardly-selected branch from each tree. These four branches were immediately transported to the laboratory for processing (detailed below). The trunk of each tree was clipped five cm above the root ball and the aboveground portion placed in a large paper bag and oven dried for 24 hrs at 60 o C; pilot experiments revealed that this length of time was sufficient to fully dry the woody biomass. We separated dry material into three tissue classes (new flush, > 1-yr needles, and woody) that were each weighed. After the aboveground portion of each tree had been removed, its root ball was carefully excavated, cleaned of all dirt and foreign objects, oven dried as above, and weighed. Belowground harvest and processing protocols are detailed elsewhere (Schaeffer et al., in prep.)

Branch-level insect densities and chemical analyses
Once the marked branches had been returned to the laboratory, we used dissecting tools to remove all of the insects without damaging any hemlock tissue.
Each branch was separated into five tissue types (new flush, 1-yr old needles, >1-yr old needles, 1-yr old stems, and >1-yr old stems) and weighed; the wet mass of each tissue type was converted to dry mass estimates using tissue-type-specific conversion factors we generated in a pilot experiment. Each tissue type was kept separate for each tree and stored at -20°C before being dried at -55°C for 72 hrs in a lyophilizer. Dried biomass from each tissue type was ground into a fine powder using a KLECO ball mill (Garcia Machines, Visalia, CA, USA).
For each marked branch, we combined the previously-collected data on insect numbers with the data on dry >1-year needle biomass to calculate branch-level insect densities (insects/gram needle). We chose this metric because (

Statistical analyses
All analyses were performed using R v. 3.2.2 (RCoreTeam 2014). We fit linear mixed effects models and used a backward-model-selection approach to examine the individual and interactive effects of HWA and EHS on hemlock. In all models, HWA and EHS were treated as fixed factors, each with three levels corresponding to the length of infestation (0, 2, or 4 years) and an interactive term (HWA*EHS). Full and reduced models were ranked and compared based on Bayesian Information Criterion (BIC) values, a standard criterion for model selection. Details of each model are contained in Appendix S2. The lme4 package was used to generate and compare models (Pinheiro et al. 2014).
In the full model, initial trunk diameter at planting was included as a covariate.
The one exception was that initial height, not initial diameter, was used as covariate for analyses of final height. Our analyses of photosynthetic rates also included time of day that the measurement was taken as a covariate. Row position (1-20) of each tree was included as a random effect in linear mixed effects models. We used this approach to examine how HWA and EHS affected the following metrics: final height, final basal diameter, total biomass, total aboveground biomass, total belowground biomass, above-/belowground biomass ratio, needle/woody biomass ratio, new flush production, photosynthesis, and bud break.
Because tissue type and age strongly impact plant chemistry, we analyzed percent C, percent N, CN ratio, total amino acids, and total starch using a slightly different approach. For stem and needle tissue, the age of the tissue (1-year or >1- year) was included in the models. Because we did not have enough tissue to conduct full suite of chemical analyses on all tissue types, analyses including new flush are limited to percent C, percent N, and CN ratio. We ran linear mixed effects models with row/harvest date as random effects.
For amino acids (AAs), we first analyzed total AA concentrations of 1-yr and >1-yr old needles using MANOVA (Wilk's λ), followed by separate two-way ANOVAs for each tissue class. We used a similar approach to examine treatment-level differences in individual AA concentrations: an initial MANOVA to account for the lack of independence amongst individual AAs, followed by two-way ANOVAs for individual amino acids. Among-treatment differences were assessed using Tukey post hoc tests. Principal component analysis was then used to compare AA composition between treatments in each tissue class.
We used linear regression to assess the relationship between branch-level insect density and the following responses of 1-year needles: photosynthesis, percent N, total AA, and total starch. Insect density was used as the predictor variable for HWA-infested trees (2, 4 years) and EHS-infested trees (2, 4 years).

Growth and biomass allocation
While HWA altered hemlock growth and biomass allocation, EHS did not.
Because the HWA*EHS interaction was never significant, we report only the main insect effects in the text (but see Appendix S2).

Early spring growth
There were no treatment-level effects of HWA or EHS on photosynthetic rates of 1-year-old foliage (Appendix S2: Table S3). Although there was no relationship between HWA density and photosynthetic rates (Appendix S2: Table S4), EHS density was negatively correlated with photosynthetic rates among trees infested with EHS for 2-and 4-years (Appendix S2: Table S5).

Foliar chemistry
While HWA substantially altered multiple aspects of foliar chemistry, EHS had virtually no impact. The N content of 1-year-old needles in HWA-infested trees, for instance, was 10% higher than in HWA-free trees (F 2,166 =10.80, P<0.001; Fig. 4a; Appendix S2: Table S7). Among trees infested with HWA for two years, HWA density was correlated with percent N; this was not, however, the case among trees infested for four years with HWA (Appendix S2: Table S4). New-flush needles on HWA-infested trees also had higher N levels (F 2,69 = 4.22, P = 0.02; Fig. 4a).
Although starch concentration in 1-year-old needles was similar in HWA-free trees and trees infested with HWA for two years, it was 70% lower in needles of trees infested with HWA for four years (Fig. 4b). EHS feeding increased starch levels in 1- year needles (F 2,155 = 3.95, P=0.02, Fig. 4b). There were no within-treatment relationships between insect density (HWA or EHS) and total starch concentration (Appendix S2: Table S4 and S5) HWA feeding increased needle AA concentrations (Wilk's λ = 0.75, P < 0.001). Total AA concentrations in 1-yr and >1-yr needles were 40% and 30% higher, respectively (both P < 0.001), than in HWA-free trees. HWA density and total AA concentrations in 1-year needles were correlated among trees infested with HWA for two years, but not trees infested with HWA for four years ( Fig. 4d; Appendix S2: Principal component analyses revealed that proline was primarily responsible for the between-treatment shifts in amino acid profiles. While HWA feeding altered the concentrations of several amino acids relative to the control, proline accounted for much of the change in total concentration (PC1=23% in one year needles, PC1=30% in greater than one year needles) (Fig. 5). While EHS did not significantly affect total amino acid concentrations, infestation caused significant changes in alanine concentration within both tissue types.

DISCUSSION
Although invasive species can play a central role in shaping temperate ecosystems and many exotic sap-feeding herbivores pose major economic and ecologic threats, we know relatively little about the long-term impact of multiple exotic herbivores. We found that chronic herbivory by two invasive sap-sucking herbivores had diverging impacts on their common host. Multiple years of HWA herbivory altered hemlock biomass allocation (above-/belowground and needle/woody), phenology, and metabolites. In contrast, EHS had minimal impacts and did not interact with HWA: dually-infested trees showed changes in allocation, phenology, and metabolites typical of HWA-only treatments (Miller-Pierce et al.
Changes in plant biomass allocation may manifest differently in distinct tissue types. Several months of HWA infestation on hemlock saplings, for instance, altered needle/woody ratios but not aboveground/belowground ratios . The needle desiccation associated with HWA feeding (Soltis et al. 2014) likely caused the needle loss among HWA-infested trees . Plants often respond to aboveground herbivory by shifting resources away from herbivore-attacked areas, often leading to an increase in belowground biomass (Babst et al. 2005, Babst et al. 2008. Given that total belowground biomass did not increase following HWA infestation, our results support previous findings ) that premature needle abscission drives changes in the above-/belowground ratio rather than mobilization of resources belowground.
Since the bulk of biomass allocation and growth in conifers occurs in early spring, open-canopy and high-light conditions early in the growing season are critical for eastern hemlock. In north-central Massachusetts, eastern hemlock forests stored the most carbon in April and May, whereas peak carbon gain in the neighboring red oak forest occurred in July and August Schedlbauer 2002, Hadley et al. 2008). Forest-atmosphere C exchange rates during the summer months in red oak forest were twice that of eastern hemlock forest (Hadley et al. 2008). The presence of HWA delayed the first day of bud break by an average of three days (Fig. 2), a period during which the overstory canopy was rapidly closing (Appendix S1: Fig. S1). This impact of this several-day delay on plant growth is exacerbated by HWA-induced reductions in new flush production (Fig. 3). Although early-April photosynthetic rates of 1-year old needles were unaffected by the presence of HWA (Fig. 4), the impact of HWA on new flush production is evident. A previous study (Gonda-King et al. 2014) showed negative impacts of HWA on photosynthesis but were completed in the fall on new foliage and thus the effect of HWA on photosynthesis likely varies seasonally.
Herbivore-attacked plants often protect themselves via induced changes in primary and secondary metabolism (Stam et al. 2014, Zhou et al. 2015. Although herbivory can alter both primary (essential functions) and secondary (in part, defense functions) metabolites, research to date has primarily addressed secondary metabolites. While changes in inducible defenses are well documented following an herbivory event, the impacts on primary metabolism are often less studied (Zhou et al. 2015). HWA and EHS feeding induced disparate effects on N metabolism, as evidenced through measures of bulk N and free amino acids.
Consistent with previous findings (Gómez et al. 2012), HWA feeding significantly elevated concentrations of N and the amino acid proline local to feeding sites. Proline accumulation is a common plant response to herbivory and drought (Delauney and Verma 1993); this and other HWA-induced changes in hemlock physiology (Radville et al. 2011, Domec et al. 2013 suggest that HWA may induce drought-like stress. In contrast, EHS feeding had no effect on either N or total amino acid concentration, regardless of tissue type or age. Interestingly, however, the responses of a suite of individual amino acids following herbivory suggests different defensive responses by hemlock to each herbivore. More specifically, HWA-infested plants had significantly lower levels of isoleucine and tryptophan. A similar pattern has been observed in Arabidopsis plants following aphid feeding, where it is associated with aphid-induced increases in the hormone abscisic acid (ABA) (Hillwig et al. 2016). Although ABA induction is often associated with water stress (Lee and Luan 2012), its induction may also benefit sap feeders via its antagonistic interactions with jasmonic acid (JA) signaling (Erb et al. 2009, Vos et al. 2013), a key pathway for anti-herbivore defense. In contrast, EHS feeding elevates alanine and phenylalanine, aromatic amino acids that are precursors for defensive plant volatiles. Phenylalanine is also an important precursor for defensive phenolic compounds (Shah 2003). The role of these defenses and signaling pathways in mediating interactions between HWA and EHS warrant further investigation.
Starch is another key primary metabolite which plays an essential role in storage. Following herbivory, stored carbohydrates are frequently remobilized.
Attacked Arabidopsis plants, for instance, increased the concentrations of enzymes employed to break down stored carbohydrates (Appel et al. 2014). The mobilization of organic compounds following attack can benefit the insect or the plant (Gómez et al. 2012, Zhou et al. 2015. Following 4-years of HWA infestation, starch in 1-year needles is reduced by approximately 30%, whereas it remains unchanged in >1-year needles (Fig. 4). Increasing EHS densities led to increases in starch in 1-year needles following 2-years of infestation, but this pattern dissipated after 4-years (Appendix S2: shady and resource-limited areas, the site described in this experiment is more realistic set of abiotic conditions and provides a full-tree perspective of the impacts of two invasive insects. Two invasive herbivores from the same feeding guild have disparate effects on biomass allocation, early spring growth, and metabolites.

Canopy monitoring
We used remotely-sensed Normalized Vegetation Difference Index (NDVI) and Leaf Area Index (LAI) values obtained from NASA MODIS (Moderate Resolution Imaging Spectroradiometer) to assess the timing and rate of forest canopy closure for the study plot in 2015. NDVI measures the difference between infrared and red wavelengths reflected off of the earth's surface, and indicates 'greenness'. LAI for broadleaf canopies is the ratio of one-sided green leaf area to ground area; it is used to indicate canopy closure. Although the measures are correlated, they measure different aspects of phenology; we looked at both to be thorough.
NDVI and LAI values for a pixel containing the study site (US-Ha1, DBF, 41.47722, 71.51166) were retrieved from the online Application for Extracting and Exploring Analysis Ready Samples (AppEEARS), courtesy of the NASA EOSDIS Land Processes Distributed Active Archive Center (LP DAAC), USGS/Earth Resources Observation and Science (EROS) Center, Sioux Falls, South Dakota, https://lpdaacsvc.cr.usgs.gov/appeears/. NDVI data were recorded at 16-day intervals for a 250 m pixel, and LAI values were calculated at 4-day intervals for a 500 m pixel. NDVI data were used if the index was produced without clouds (VI Quality MODLAND = 00 or 01) and pixel usefulness was 'highest' or 'lower' (VI usefulness = 0000 or 0001). LAI data were used if the pixel was 'good quality' (i.e. MODLAND = 0) and the cloud state was 'clear' or 'mixed' (i.e. CloudState ≠ 01). Figure S1. Results of models assessing canopy closure at the field site in early spring. Vertical lines show the averaged bud break day among un-infested (pink), EHS-infested (lime green), and HWA-infested (blue) trees. Table S1. Results of linear mixed effect models ( †) assessing the impacts of HWA and EHS on full tree metrics. When linear mixed model (denoted with †) were used, row was used as the random effect. In all cases we report Type II error. Table S2. Results of ANOVAs (¢) assessing the impacts of HWA and EHS on biomass ratios. In all cases we report Type II error. Table S4. HWA density (insects/gram) against photosynthetic rate, Percent N, total starch, and total AA in 1-year needles. All statistics were completed with linear models. Models are split by length of HWA infestation (2-years of HWA and 4-years of HWA). These groups include dually infested trees. A dash indicates that the metric was not used in the final model. Table S5. EHS density (insects/gram) against photosynthetic rate, Percent N, total starch, and total AA in 1-year needles. All statistics were completed with linear models. Models are split by length of EHS infestation (2-years EHS and 4-years EHS). These groups include dually infested trees. A dash indicates that the metric was not used in the final model. Table S6. Results of a linear mixed effects model ( †) and ANOVA (¢) assessing the impacts of HWA and EHS on bud break and daily new flush production in mid-May. Table S8. Results from Linear Mixed Effects Models assessing the impact of HWA, EHS, and foliage age on total AA and total starch among needles and stems tissue.