Publications Library

Given the coarse resolution of global climate models, downscaling techniques are often needed to generate finer scale projections of variables affected by local-scale processes such as precipitation. However, classical statistical downscaling experiments for future climate rely on the time-invariance assumption as one cannot know the true change in the variable of interest, nor validate the models with data not yet observed. Our experimental setup involves using the Canadian regional climate model (CRCM) outputs as pseudo-observations to estimate model performance in the context of future climate projections by replacing historical and future observations with model simulations from the CRCM, nested within the domain of the Canadian global climate model (CGCM). In particular, we evaluated statistically downscaled daily precipitation time series in terms of the Peirce skill score, mean absolute errors, and climate indices. Specifically, we used a variety of linear and nonlinear methods such as artificial neural networks (ANN), decision trees and ensembles, multiple linear regression, and k-nearest neighbors to generate present and future daily precipitation occurrences and amounts. We obtained the predictors from the CGCM 3.1 20C3M (1971–2000) and A2 (2041–2070) simulations, and precipitation outputs from the CRCM 4.2 (forced with the CGCM 3.1 boundary conditions) as predictands. Overall, ANN models and tree ensembles outscored the linear models and simple nonlinear models in terms of precipitation occurrences, without performance deteriorating in future climate. In contrast, for the precipitation amounts and related climate indices, the performance of downscaling models deteriorated in future climate.

Fully coupled global climate model experiments are performed using the Community Climate System Model version 4.0 (CCSM4) for preindustrial, present, and future climate to study the effects of realistic land surface initializations on subseasonal to seasonal climate forecasts. Model forecasts are verified against model control simulations (perfect model experiments), thus overcoming to some extent issues of uncertainties in the observations and/or model parameterizations. Findings suggest that realistic land surface initialization is important for climate predictability at subseasonal to seasonal time scales. We found the highest predictability for soil moisture, followed by evapotranspiration, temperature, and precipitation. The predictability is highest for the 16 to 30 days forecast period, and it progressively decreases for the second and third month forecasts. We found significant changes in the spatial distributions of temperature predictability in the present and future climate compared to the preindustrial climate, although the spatial average changes for North America were rather small (

Regional warming due to anthropogenic influence on the climate is expected to increase the frequency of very warm years and seasons. The growing research area of extreme event attribution has provided pertinent scientific evidence for a number of such warm events for which the forced climate response rises above internal climatic variability. Although the demand for attribution assessments is higher shortly after an event occurs, most scientific studies become available several months later. A formal attribution methodology is employed here to pre-compute the changing odds of very warm years and seasons in regions across the world. Events are defined based on the exceedence of temperature thresholds and their changing odds are measured over a range of pre-specified thresholds, which means assessments can be made as soon as a new event happens. Optimal fingerprinting provides observationally constrained estimates of the global temperature response to external forcings from which regional information is extracted. This information is combined with estimates of internal variability to construct temperature distributions with and without the effect of anthropogenic influence. The likelihood of an event is computed for each distribution and the change in the odds estimated. Analyses are conducted with seven climate models to explore the model dependency of the results. Apart from colder regions and seasons, characterised by greater internal climate variability, the odds of warm events are found to have significantly increased and temperatures above the threshold of 1-in-10 year events during 1961–1990 have become at least twice as likely to occur.

The discrepancy between recent observed and simulated trends in global mean surface temperature has provoked a debate about possible causes and implications for future climate change projections. However, little has been said in this discussion about observed and simulated trends in global temperature extremes. Here we assess trend patterns in temperature extremes and evaluate the consistency between observed and simulated temperature extremes over the past four decades (1971–2010) in comparison to the recent 15 years (1996–2010). We consider the coldest night and warmest day in a year in the observational dataset HadEX2 and in the current generation of global climate models (CMIP5). In general, the observed trends fall within the simulated range of trends, with better consistency for the longer period. Spatial trend patterns differ for the warm and cold extremes, with the warm extremes showing continuous positive trends across the globe and the cold extremes exhibiting a coherent cooling pattern across the Northern Hemisphere mid-latitudes that has emerged in the recent 15 years and is not reproduced by the models. This regional inconsistency between models and observations might be a key to understanding the recent hiatus in global mean temperature warming.

Logistical constraints can limit the number of Global Climate Model (GCM) simulations considered in a climate change impact assessment. When dealing with annual or seasonal variables, one can visualize and manually select GCM scenarios to cover as much of the ensemble’s range of changes as possible. Most environmental systems are sensitive to climate conditions, e.g., extremes, that cannot be described by a small number of variables. Instead, algorithms like k-means clustering have been used to select representative ensemble members. Clustering algorithms are, however, biased towards high-density regions of climate variable space and tend to select scenarios that describe the central tendency rather than the full spread of an ensemble. Also, scenarios selected via clustering may not be ordered, i.e., scenarios in the 5 cluster solution may not appear in the 6 cluster solution, which makes recommending a consistent set of scenarios to researchers with different needs difficult. Alternatively, an automated procedure based on a cluster initialization algorithm is proposed and applied to changes in 27 climate extremes indices between 1986-2005 and 2081-2100 from a large ensemble of Coupled Model Intercomparison Project Phase 5 (CMIP5) simulations. Selections by the method are ordered and are designed to span the overall range of the ensemble. The number of scenarios required to account for changes spanned by at least 90% of the CMIP5 ensemble members is reported for 21 regions of the globe and compared with k-means clustering. On average, the proposed method requires 40% fewer scenarios to meet this threshold than does k-means clustering.

The current distribution of forest genetic resources on Earth is the result of a combination of natural processes and human actions. Over time, tree populations have become adapted to their habitats including the local ecological disturbances they face. As the planet enters a phase of human-induced climate change of unprecedented speed and magnitude, however, previously locally-adapted populations are rendered less suitable for new conditions, and ‘natural’ biotic and abiotic disturbances are taken outside their historic distribution, frequency and intensity ranges. Tree populations rely on phenotypic plasticity to survive in extant locations, on genetic adaptation to modify their local phenotypic optimum or on migration to new suitable environmental conditions. The rate of required change, however, may outpace the ability to respond, and tree species and populations may become locally extinct after specific, but as yet unknown and unquantified, tipping points are reached. Here, we review the importance of forest genetic resources as a source of evolutionary potential for adaptation to changes in climate and other ecological factors. We particularly consider climate-related responses in the context of linkages to disturbances such as pests, diseases and fire, and associated feedback loops. The importance of management strategies to conserve evolutionary potential is emphasised and recommendations for policy-makers are provided.

Regional warming due to anthropogenic influence on the climate is expected to increase the frequency of very warm years and seasons. The growing research area of extreme event attribution has provided pertinent scientific evidence for a number of such warm events for which the forced climate response rises above internal climatic variability. Although the demand for attribution assessments is higher shortly after an event occurs, most scientific studies become available several months later. A formal attribution methodology is employed here to pre-compute the changing odds of very warm years and seasons in regions across the world. Events are defined based on the exceedence of temperature thresholds and their changing odds are measured over a range of pre-specified thresholds, which means assessments can be made as soon as a new event happens. Optimal fingerprinting provides observationally constrained estimates of the global temperature response to external forcings from which regional information is extracted. This information is combined with estimates of internal variability to construct temperature distributions with and without the effect of anthropogenic influence. The likelihood of an event is computed for each distribution and the change in the odds estimated. Analyses are conducted with seven climate models to explore the model dependency of the results. Apart from colder regions and seasons, characterised by greater internal climate variability, the odds of warm events are found to have significantly increased and temperatures above the threshold of 1-in-10 year events during 1961–1990 have become at least twice as likely to occur.

A recent hydrological impacts study in British Columbia, Canada, used an ensemble of 23 climate change simulations to assess potential future changes in streamflow. These Coupled Model Intercomparison Project Phase 3 (CMIP3) simulations were statistically downscaled and used to drive the Variable Infiltration Capacity (VIC) hydrology model over several watersheds. Due to computational restrictions, the 23 member VIC ensemble is a subset of the full 136 member CMIP3 archive. Extending the VIC ensemble to cover the full range of uncertainty represented by CMIP3, and incorporating the latest generation CMIP5 ensembles, poses a considerable computing challenge. Thus, we extend the VIC ensemble using a computationally efficient statistical emulation model, which approximates the combined output of the two-step process of statistical downscaling and hydrologic modeling, trained with the 23 member VIC ensemble. Regularized multiple linear regression links projected changes in monthly temperature and precipitation with projected changes in monthly streamflow over the Fraser and Peace River watersheds. Following validation, the statistical emulator is forced with the full suite of CMIP3 and CMIP5 climate change projections. The 23 member VIC ensemble has a smaller spread than the full ensemble; however, both ensembles provide the same consensus estimate of monthly streamflow change. Qualitatively, CMIP5 shows a similar streamflow response as CMIP3 for snow-dominated hydrologic regimes. However, by end-century, the CMIP5 worst-case RCP8.5 has a larger impact than CMIP3 A2. This work also underscores the advantage of using emulation to rapidly identify those future extreme projections that may merit further study using more computationally demanding process-based methods.

Attribution of extreme events is a challenging science and one that is currently undergoing considerable evolution. In this paper, 20 different research groups explored the causes of 16 different events that occurred in 2013. The findings indicate that human-caused climate change greatly increased the risk for he extreme heat waves assessed in this report. How human influence affected other types of events such as droughts, heavy rain events, and storms was less clear, indicating that natural variability likely played a much larger role in these extremes. Multiple groups chose to look at both the Australian heat waves and the California drought, providing an opportunity to compare and contrast the strengths and weaknesses of various methodologies. There was considerable agreement about the role anthropogenic climate change played in the events between the different assessments. This year three analyses were f evere storms and none found an anthropogenic signal. However, attribution assessments of these types of events pose unique challenges due to the often limited observational record. When human-influence for an event is not identified with the scientific tools available to us today, this means that if there is a human contribution, it cannot be distinguished
from natural climate variability.