Dendrochronology is a technique that can be applied as a retrospective monitoring (RM) approach to evaluate the performance of nutrient application in forest ecosystems. Applying the RM approach across operations lacks experimental controls, which may adversely affect accuracy and precision of estimates due to greater mismatches in stand conditions between treated and untreated plots. To test the rigor of the RM approach, we collected increment cores of coastal Douglas fir (Pseudotsuga menziesii var. menziesii) at eight sites of an experiment where stands were fertilized in 1971. First, we tested the approach under ideal conditions by sampling from treated and untreated plots of the experiment. Second, we tested the approach using newly established surrogate control (SC) plots, which differed in ecological site classification from those of the treated plots to understand how robust the approach was to mismatches in conditions between treated and untreated samples. We hypothesized that detrending ring width would mitigate error in responses to nutrient application resulting from mismatches in site classification. Within the experiment, the approach indicated an average increase of 15% (5% to 26%, p < 0.05) growth response to operational doses of urea. Different responses were found when the analysis relied on SC plots. Detrending low-frequency variation in ring widths eliminated differences in results arising, at least in part, to mismatches in site class. However, it also reduced the growth response using the experimental control plots to 10%. Dendrochronology with detrending shows promise in the ability to mitigate variation introduced by mismatches in ecological site classification that may occur in operational monitoring. Based on these results, we see potential to implement RM with operations to evaluate and optimize stand selection criteria.

The claim that changes in atmospheric composition and climate have enhanced the growth rate of trees is prevalent in science, yet it is not supported by many recent tree ring studies. In this study, we analyzed historical time trends in stemwood biomass growth derived from black spruce (Picea mariana (Mill.) BSP) trees at 248 plots across Canada. The sample consisted of trees that were live and dominant at the time of sampling (LDS). Observations of stemwood biomass of LDS trees at a reference age of 75 years (B-sw75,B-LDS) increased by 154 to 321% over 1901-2001 depending on the method of trend estimation. Simulations from a calibrated individual-based Growth and Yield model-forced with varying degrees of hypothetical trend in tree growth-were used to estimate the proportion of trend that could be attributed to intrinsic factors, including artefacts introduced by sampling from LDS trees instead of from the population of trees. Imposing no time trend in simulated tree growth, stemwood biomass of 75-year-old LDS trees (B-sw75,B-LDS,B-Model) increased by 63% (41 to 85% CI). We conclude that the remaining variation in growth of LDS trees (154 to 321% minus 63% = 91 to 258%) can be attributed to net extrinsic forcing. The scaling relationship between LDS and population trees further suggested that stemwood biomass growth of the population (G(sw75,POP)) increased by 47 to 82%. By accounting for both natural dynamics and artefacts of the sampling design in estimation of net intrinsic forcing, we gained confidence that growth rate of black spruce trees across Canada increased significantly over 1901-2001. While growth enhancement is consistent with beneficial effects of increasing levels of reactive nitrogen, carbon dioxide, and warming, there remains uncertainty in the degree that applied procedures fully account for known sampling artefacts.

This report describes the North American Climate Integration and Diagnostics – Nitrogen Deposition Version 1 (NACID-NDEP1) database. The database contains estimates of annual total nitrogen (N) deposition for the purpose of supporting terrestrial ecosystem modelling in North America. It was constructed at 1-km resolution with coverage of Alaska, Canada, and the conterminous U.S., with continuous annual coverage from 1860 to 2013. Estimates were produced by acquiring and compiling best-available data sources: Wet N deposition was estimated from interpolation of monthly ammonium and nitrate concentration measurements and from grids of monthly precipitation. Dry N deposition was estimated from satellite measurements of ammonium and nitrogen oxides. Total N deposition for the pre-industrial era was derived from previous modelling studies. As these source datasets covered different time periods, several assumptions were required to produce a continuous record.

Environmental conditions and silviculture fundamentally alter the metabolism of individual trees and, therefore, need to be studied at that scale. However, changes in forest biomass density (Mg C ha−1) may be decoupled from changes in growth (kg C year−1) when the latter also accelerates the life cycle of trees and strains access to light, nutrients, and water. In this study, we refer to an individual-based model of forest biomass dynamics to constrain the magnitude of system feedbacks associated with ontogeny and competition and estimate the scaling relationship between changes in tree growth and forest biomass density. The model was driven by fitted equations of annual aboveground biomass growth (Gag), probability of recruitment (Pr), and probability of mortality (Pm) parameterized against field observations of black spruce (Picea mariana (Mill.) BSP), interior Douglas-fir (Pseudotsuga menziesii var. glauca (Beissn.) Franco), and western hemlock (Tsuga heterophylla (Raf.) Sarg.). A hypothetical positive step-change in mean tree growth was imposed half way through the simulations and landscape-scale responses were then evaluated by comparing pre- and post-stimulus periods. Imposing a 100% increase in tree growth above calibrated predictions (i.e., contemporary rates) only translated into 36% to 41% increases in forest biomass density. This corresponded with a tree-growth elasticity of forest biomass (εG,SB) ranging from 0.33 to 0.55. The inelastic nature of stand biomass density was attributed to the dependence of mortality on intensity of competition and tree size, which decreased stand density by 353 to 495 trees ha−1, and decreased biomass residence time by 10 to 23 years. Values of εG,SB depended on the magnitude of the stimulus. For example, a retrospective scenario in which tree growth increased from 50% below contemporary rates up to contemporary rates indicated values of εG,SB ranging from 0.66 to 0.75. We conclude that: (1) effects of warming and increasing atmospheric concentrations of carbon dioxide and reactive nitrogen on biomass production are greatly diminished, but not entirely precluded, scaling up from individual trees to forest landscapes; (2) the magnitude of decoupling is greater for a contemporary baseline than it is for a pre-industrial baseline; and (3) differences in the magnitude of decoupling among species were relatively small. To advance beyond these estimates, studies must test the unverified assumptions that effects of tree size and stand competition on rates of recruitment, mortality, and growth are independent of climate change and atmospheric concentrations of carbon dioxide and nitrogen.

Given the widespread ecological implications that would accompany any significant change in evaporative demand of the atmosphere, this study investigated spatial and temporal variation in several accepted expressions of potential evaporation (PE). The study focussed on forest regions of North America, with 1 km-resolution spatial coverage and a monthly time step, from 1951–2014. We considered Penman's model (EPen), the Priestley–Taylor model (EPT), 'reference' rates based on the Penman–Monteith model for grasslands (ERG), and reference rates for forests that are moderately coupled (ERFu) and well coupled (ERFc) to the atmosphere. To give context to the models, we also considered a statistical fit (EPanFit) to measurements of pan evaporation (EPan). We documented how each model compared with EPan, differences in attribution of variance in PE to specific driving factors, mean spatial patterns, and time trends from 1951–2014. The models did not agree strongly on the sensitivity to underlying drivers, zonal variation of PE, or on the magnitude of trends from 1951–2014. Sensitivity to vapour pressure deficit (Da) differed among models, being absent from EPT and strongest in ERFc. Time trends in reference rates derived from the Penman–Monteith equation were highly sensitive to how aerodynamic conductance was set. To the extent that EPanFit accurately reflects the sensitivity of PE to Da over land surfaces, future trends in PE based on the Priestley–Taylor model may underestimate increasing evaporative demand, while reference rates for forests, that assume strong canopy-atmosphere coupling in the Penman–Monteith model, may overestimate increasing evaporative demand. The resulting historical database, covering the spectrum of different models of PE applied in modern studies, can serve to further investigate biosphere-hydroclimate relationships across North America.

Accounting for water stress-induced tree mortality in forest productivity models remains a challenge due to uncertainty in stress tolerance of tree populations. In this study, logistic regression models were developed to assess species- specific relationships between probability of mortality (P-m) and drought, drawing on 8.1 million observations of change in vital status (m) of individual trees across North America. Drought was defined by standardized (relative) values of soil water content (W-s,W- z) and reference evapotranspiration (ETr, z) at each field plot. The models additionally tested for interactions between the water-balance variables, aridity class of the site (AC), and estimated tree height (h). Considering drought improved model performance in 95 (80) per cent of the 64 tested species during calibration (cross-validation). On average, sensitivity to relative drought increased with site AC (i.e. aridity). Interaction between water-balance variables and estimated tree height indicated that drought sensitivity commonly decreased during early height development and increased during late height development, which may reflect expansion of the root system and decreasing whole-plant, leaf-specific hydraulic conductance, respectively. Across North America, predictions suggested that changes in the water balance caused mortality to increase from 1.1% yr(-1) in 1951 to 2.0% yr(-1) in 2014 (a net change of 0.9 +/- 0.3% yr(-1)). Interannual variation in mortality also increased, driven by increasingly severe droughts in 1988, 1998, 2006, 2007 and 2012. With strong confidence, this study indicates that water stress is a common cause of tree mortality. With weak-to-moderate confidence, this study strengthens previous claims attributing positive trends in mortality to increasing levels of water stress. This 'learn-as-we-go' approach - defined by sampling rare drought events as they continue to intensify - will help to constrain the hydraulic limits of dominant tree species and the viability of boreal and temperate forest biomes under continued climate change.

New initiatives are emerging to rehabilitate vast expanses of lodgepole pine (Pinus contorta var. latifolia) recently affected by insect outbreaks and wildfires across western North America. By assessing potential long-term changes in productivity, environmentally-sensitive models of tree growth can help to make informed decisions about where lodgepole pine will likely thrive and how it will respond to silvicultural treatments under future environmental conditions. It remains a challenge, however, to produce models that are practical to apply, statistically robust, and representative of real-world complexity. In this study, we developed composite models of aboveground biomass growth of lodgepole pine, with the purpose of assessing the statistical performance and behaviour of models covering a range of complexity. Adding complexity, including environmental sensitivity, did not appreciably increase explained variance over that of models based on tree size, stand competition, and site conditions alone. Nevertheless, Akaike's Information Criterion and cross-validation results offered no reason to dismiss the most complex model, which additionally considered insects, pathogens, heat, water, nitrogen deposition, and carbon dioxide concentration, and interactions among covariates. For models that represented environmental sensitivity, average predictions of growth increased between historical and future periods, but the increases were not significant at the 95% confidence level. Despite consistency in the overall average future change in growth among environmentally-sensitive models, the consideration of interactions among environmental variables caused fundamental changes in the magnitude and spatial pattern of future predictions, suggesting that interactive responses of tree growth to changing climate and atmospheric composition remain an important knowledge gap, that can likely only be constrained by improved coordination of field sampling and experimentation.

Repeated measurements of tree biomass at field plots describe recovery from disturbances, sampling artifacts, and potential effects of environmental change on forest ecosystems. Challenges in differentiating between intrinsic and extrinsic sources of variation, both in theory and in practice, continue to confound claims of an anthropogenic carbon sink in forest biomass. Here we analyzed observations at 10,307 plots across southern ecozones of Canada to investigate temporal trends in stand‐level biomass growth (G), biomass loss due to mortality (M), and net ecosystem biomass production (NEBP) of intact stands. Net extrinsic forcing (Fex) was expressed by the collective dependence of biomass fluxes on climate anomalies, nitrogen deposition (N), and atmospheric carbon dioxide concentration (C). Inferences drawn directly from linear mixed‐effects model coefficients only reflect the static behavior of the model specifically at field plot locations. We, therefore, defined a dynamic landscape‐scale net extrinsic forcing (Fex′), which additionally accounted for potential negative feedback responses to anthropogenic growth enhancement. Simulations were performed over 1501–2012 to estimate Fex′. Overall, Fex′ was positive, suggesting that environmental changes drove a 90% increase in NEBP. The increase in NEBP was confined to wet regions, while the biomass sink in dry regions decreased, suggesting that large expanses of northern forests, historically located near the boundary between wet and dry climates, may be at high risk of decline under continued increases in evaporative demand. These results have important implications for the greenhouse gas balance of Canada's forest sector.

The impacts of climate change and increasing atmospheric CO2 concentration on the terrestrial uptake of carbon dioxide since 1860 in the Canadian province of British Columbia are estimated using the process-based Canadian Terrestrial Ecosystem Model (CTEM). Model simulations show that these two factors yield an enhanced carbon uptake of around 44 gC m−2 yr−1 (or equivalently 63 gC m−2 yr−1 over the province's forested area), during the 1980s and 1990s, and continuing into the 2000s. About three-quarters of the simulated sink enhancement in our study compared to pre-industrial conditions is attributed to changing climate, and the rest is attributed to increase in CO2 concentration. The model response to changing climate and increasing CO2 is corroborated by comparing simulated stem wood growth rates with ground-based measurements from inventory plots in coastal British Columbia. The simulated sink is not an estimate of the net carbon balance because the effects of harvesting, insect disturbances and land-use change are not considered.

Drought-related dieback of aspen-dominated woodland has been a persistent and possibly increasing phenomenon over recent decades in the southern boreal forests of western Canada. The Integrated Biosphere Simulator (IBIS) dynamic vegetation model was modified for Canadian ecosystems (hence "Can-IBIS") and used to simulate effects of water stress on the woody biomass of aspen-dominated stands in the boreal mixedwood regions of Saskatchewan and Alberta. The modified model was evaluated using eddy-covariance measurements of CO2 and water vapor fluxes made at a forested site and a grassland site located in the study region. The tested model captured 74% of the variation in biomass growth trajectories at 13 boreal and 12 parkland field study sites; the mean difference between simulated and observed values was approximately 1100 g C m(-2). Under the combined influences of climatic variation and increasing atmospheric CO2 from 2000 to 2008, simulated values of net biomass growth were 544 and 240 g C m(-2) for the boreal and parkland study sites, respectively. Drought-induced biomass losses at the drier sites (in both boreal and parkland regions) were simulated to be 100-350 g C m(-2), corresponding to an annual modeled mortality rate of 5-7% during severe drought years. These results were consistent with field measurements and other statistical studies. Changes in biomass over the nine-year period varied with geographical location and corresponded to spatial variation in monthly values of the self-calibrated Palmer Drought Severity Index. We conclude that Can-IBIS can be used to investigate annual impacts of water stress on woody biomass growth, although cumulative physiological effects of multi-year droughts on tree mortality would benefit from improved simulation of subgrid-scale (soil texture-driven) processes. In particular, two areas for further development are: (1) calibration based on the results of soil surveys at fine spatial/temporal scales; and (2) biophysical experiments to refine the representation of water stress constraints on biomass turnover. (C) 2014 Elsevier B.V. All rights reserved.

Changes in climate and atmospheric CO2 and nitrogen (N) over the last several decades have induced significant effects on forest carbon (C) cycling. However, contributions of individual factors are largely unknown because of the lack of long observational data and the undifferentiating between intrinsic factors and external forces in current ecosystem models. Using over four decades (1956–2001) of forest inventory data at 3432 permanent samples in maritime and boreal regions of British Columbia (B.C.), Canada, growth enhancements were reconstructed and partitioned into contributions of climate, CO2 and N after removal of age effects. We found that climate change contributed a particularly large amount (over 70%) of the accumulated growth enhancement, while the remaining was attributed to CO2 and N, respectively. We suggest that climate warming is contributing a widespread growth enhancement in B.C.'s forests, but ecosystem models should consider CO2 and N fertilization effects to fully explain inventory-based observations.

We assess Canada's gross primary productivity (GPP) and net primary productivity (NPP) using boreal ecosystem productivity simulator (BEPS) at 250m spatial resolution with improved input parameter and driver fields and phenology and nutrient release parameterization schemes. BEPS is a process-based two-leaf enzyme kinetic terrestrial ecosystem model designed to simulate energy, water, and carbon (C) fluxes using spatial data sets of meteorology, remotely sensed land surface variables, soil properties, and photosynthesis and respiration rate parameters. Two improved key land surface variables, leaf area index (LAI) and land cover type, are derived at 250m from Moderate Resolution Imaging Spectroradiometer sensor. For diagnostic error assessment, we use nine forest flux tower sites where all measured C flux, meteorology, and ancillary data sets are available. The errors due to input drivers and parameters are then independently corrected for Canada-wide GPP and NPP simulations. The optimized LAI use, for example, reduced the absolute bias in GPP from 20.7% to 1.1% for hourly BEPS simulations. Following the error diagnostics and corrections, daily GPP and NPP are simulated over Canada at 250m spatial resolution, the highest resolution simulation yet for the country or any other comparable region. Total NPP (GPP) for Canada's land area was 1.27 (2.68) Pg C for 2008, with forests contributing 1.02 (2.2) Pg C. The annual comparisons between measured and simulated GPP show that the mean differences are not statistically significant (p>0.05, paired t test). The main BEPS simulation error sources are from the driver fields.

The forest carbon cycle is not only controlled by climate, tree species and site conditions, but also by disturbance affecting the biomass and age of forest stands. The Carbon Budget Model of the Canadian forest sector (CBM-CFS3) calculates the complete forest carbon cycle by combining forest inventory data on forest species, biomass and stand age with empirical yield information and statistics on forest disturbances, management and land-use change. It is used for national reporting and climate policy purposes. The Integrated Terrestrial Ecosystem Carbon model (InTEC) is driven by remotely-sensed vegetation parameters (forest type, leaf area index, clumping index) and fire scar, soil and climate data and simulates forest growth and the carbon cycle as a function of stand age using a process-based approach. Gridded forest biomass, stand age and disturbance data based on forest inventory are also used as …

Department of Renewable Resources, University of Alberta, Edmonton, AB, Canada T6G 2E3 School of Geography and Earth Sciences and McMaster Centre for Climate Change, McMaster University, Hamilton, Ont., Canada Department of Forest Resources Management, University of British Columbia, Vancouver, BC, Canada Natural Resources Canada, Canadian Forest Service, Victoria, BC, Canada Natural Resources Canada, Canadian Forest Service, Edmonton, Alta., Canada

To understand how environmental changes have influenced forest productivity, stemwood biomass ( B ) dynamics were analyzed at 1267 permanent inventory plots, covering a combined 209 ha area of unmanaged temperate‐maritime forest in southwest British Columbia, Canada. Net stemwood production (Δ B ) was derived from periodic remeasurements of B collected over a 40‐year measurement period (1959–1998) in stands ranging from 20 to 150 years old. Comparison between the integrated age response of net stemwood production, Δ B ( A ), and the age response of stemwood biomass, B ( A ), suggested a 58 ± 11% increase in Δ B between the first 40 years of the chronosequence period (1859–1898) and the measurement period. To estimate extrinsic forcing on Δ B , several different candidate models were developed to remove variation explained by intrinsic factors. All models exhibited temporal bias, with positive trends in (observed minus predicted) residual Δ B ranging between of 0.40 and 0.64% yr −1 . Applying the same methods to stemwood growth ( G ) indicated residual increases ranging from 0.43 and 0.67% yr −1 . Higher trend estimates corresponded with models that included site index ( SI ) as a predictor, which may reflect exaggeration of the age‐decline in SI tables. Choosing a model that excluded SI , suggested that Δ B increased by 0.40 ± 0.18% yr −1 , while G increased by 0.43 ± 0.12% yr −1 over the measurement period. Residual G was significantly correlated with atmospheric carbon dioxide ( CO 2 ), temperature ( T ), and climate moisture index ( CMI ). However, models driven with climate and CO 2 , alone, could not simultaneously explain long‐term and measurement‐period trends without additional representation of indirect effects, perhaps reflecting compound interest on direct physiological responses to environmental change. Evidence of accelerating forest regrowth highlights the value of permanent inventories to detect and understand systematic changes in forest productivity caused by environmental change.

Forest productivity is strongly affected by seasonal weather patterns and by natural or anthropogenic disturbances. However weather effects on forest productivity are not currently represented in inventory-based models such as CBM-CFS3 used in national forest C accounting programs. To evaluate different approaches to modelling these effects, a model intercomparison was conducted among CBM-CFS3 and four process models (ecosys, CN-CLASS, Can-IBIS and 3PG) over a 2500 ha landscape in the Oyster River (OR) area of British Columbia, Canada. The process models used local weather data to simulate net primary productivity (NPP), net ecosystem productivity (NEP) and net biome productivity (NBP) from 1920 to 2005. Other inputs used by the process and inventory models were generated from soil, land cover and disturbance records. During a period of intense disturbance from 1928 to 1943, simulated NBP diverged considerably among the models. This divergence was attributed to differences among models in the sizes of detrital and humus C stocks in different soil layers to which a uniform set of soil C transformation coefficients was applied during disturbances. After the disturbance period, divergence in modelled NBP among models was much smaller, and attributed mainly to differences in simulated NPP caused by different approaches to modelling weather effects on productivity. In spite of these differences, age-detrended variation in annual NPP and NEP of closed canopy forest stands was negatively correlated with mean daily maximum air temperature during July-September (T-amax) in all process models (R-2 = 0.4-0.6), indicating that these correlations were robust. The negative correlation between T-amax. and NEP was attributed to different processes in different models, which were tested by comparing CO2 fluxes from these models with those measured by eddy covariance (EC) under contrasting air temperatures (T-a). The general agreement in sensitivity of annual NPP to T-amax, among the process models led to the development of a generalized algorithm for weather effects on NPP of coastal temperate coniferous forests for use in inventory-based models such as CBM-CFS3: NPP' = NPP - 57.1 (T-amax - 18.6), where NPP and NPP' are the current and temperature-adjusted annual NPP estimates from the inventory-based model, 18.6 is the long-term mean daily maximum air temperature during July-September, and T-amax is the mean value for the current year. Our analysis indicated that the sensitivity of NPP to T-amax was nonlinear, so that this algorithm should not be extrapolated beyond the conditions of this study. However the process-based methodology to estimate weather effects on NPP and NEP developed in this study is widely applicable to other forest types and may be adopted for other inventory based forest carbon cycle models. (C) 2011 Elsevier B.V. All rights reserved.

Eddy-covariance (EC) measurements in three different-aged Douglas-fir (Pseudotsuga menziesii [Mirb.] Franco var. menziesii) stands were used to calibrate a production efficiency model (PEM) and explore the sources of error in simulated annual gross primary production (Pg). Parameters were derived on a daily time scale, assessing absorbed photosynthetically active radiation (Qa), maximum gross photosynthetic efficiency (ɛg max), and functions of environmental stress. Despite similar climate, ɛg max varied between sites in correspondence with measurements of site index derived from forest inventory, suggesting that landscape variation of ɛg max is controlled mainly by non-climatic factors and ranges approximately between 1.28 and 4.42gCMJ−1. Within stands, daily variation of Pg was most strongly controlled by decreasing ɛg with increasing Qa. We therefore devised a method of incorporating the nonlinear light response (NLR) that is apparent within stands into the model, while preserving the linearity in the relationship between annual Pg and Qa across stands that is assumed in conventional PEMs. The ability to match observed seasonal and inter-annual variability of Pg improved by taking into account antecedent effects of cumulative heat on plant development. An ecosystem-specific model (i.e., fitted collectively to all stands) explained 81, 95, and 97% of the monthly variation of Pg in regenerating, juvenile, and mature stands, respectively. The model was able to collectively explain 96% of variability of annual total Pg with a root mean squared error 130gCm−2yr−1, constituting 6% of the mean and 113% of the standard deviation at the mature site. The capacity to predict inter-annual variability of Pg was strongly limited by discrepancies that persisted for days-to-weeks at a time, which implies that poor model skill was caused, to uncertain degrees, by inadequate representation of acclimation to environmental stress, and discrepancies between measurement and footprint-weighted conditions.