No description available.

No description available.

Improved characterization of cloud and precipitation features are required to understand the impact of a changing climate in high latitude regions and accurately represent these features in models. The importance of cold season precipitation to regional moisture cycling and our limited understanding of orographic cloud and precipitation processes in the Arctic provide the motivation for this research. Using high-resolution datasets collected during the Storm Studies in the Arctic (STAR) field project this thesis examines cloud and precipitation features over southern Baffin Island in Nunavut. Cloud and precipitation features were shown to differ over orography compared to the adjacent ocean regions upstream. Gravity waves, terrain shape, atmospheric stability and atmosphere-ocean exchanges were all associated with precipitation enhancement. In addition, high sea ice extent, low-level blocking in the upstream environment and sublimation were factors that reduced precipitation. The nature of hydrometeors was variable and accretion and aggregation were found to be important determinants of whether precipitation reached the ground. The processes controlling a snowfall event over southern Baffin Island were found to be complex, representing a significant challenge for modelling in the region. Low-level convection over adjacent ocean regions, strong upslope flow over the terrain, and the passing of a weak trough collectively produced the event. Analysis of the Global Environmental Multi-scale limited area model (GEM-LAM 2.5) revealed that upstream convection and upslope processes were affected by model errors. Consequently, precipitation onset was delayed and total modelled accumulation was 50% less than observations. Further evaluation of a numerical weather prediction model during STAR cases provided descriptions of model errors and proficiencies for different synoptic forcing and surface environments. Overall the model overestimated temperature and had difficulties representing thermal inversions over sea ice. The model generally over-predicted moisture with the exception of profiles over sea ice and land. Wind speed was frequently underestimated, weakening upslope processes and errors in wind direction were large at times. Cloud-tops were usually too high and cloud-bases too low. Where multiple cloud layers were present, the dry layer depth was inaccurate. Model errors were shown to have implications for cloud and precipitation production and their forecast.

Stochastic generation of daily precipitation at multiple sites is frequently needed to evaluate the long-term effects of hydrologic and climate-change in design and operation of water resources systems. Capturing the spatial dependence of precipitation at multiple sites into a stochastic model presents a great challenge because of the non-normal bivariate distributions of precipitation-amounts. Without normalizing the precipitation amounts, many models have attempted to establish spatial dependence through alternative methods that tended to be cumbersome. In contrast, representing precipitation in Gaussian fields provides a generic structure that is well-amenable to statistical analyses facilitating easy implementation of models. The thrust of this thesis is to generate normalized precipitation data and transform them back into the original domain for applications and analyses. A multivariate censored distribution (MCD) and a multivariate autoregressive censored process (MACP) are developed to formulate two weather generation (WG) models. Parameters of censored distributions were estimated by using the maximum likelihood method. To reduce the magnanimity in the number of parameters and their temporal variation, elements of covariance matrices of models were represented by periodic functions. The performance of models was evaluated by comparing discrepancies in attributes. Three performance measures (i.e., the coefficient of determination, the coefficient efficiency and the root mean square error) suggested that simulated data to be indistinguishable from the historical precipitation sequences. The models were implemented with other techniques to address the three most common problems encountered in daily precipitation records. The first implementation is related to simulation of precipitation at un-gauged sites using the WG-MACP model with general regression neural networks or Kriging methods. The second implementation was related to infilling of missing observations a using the WG-MCD and WG-MACP models with Gibbs sampling. The third implementation was related to downscaling of monthly and daily output of the Canadian regional climate model (CRCM) using traditional and parametric Delta change methods.

The Canadian Precipitation Analysis (CaPA) generates gridded precipitation data outputs based on the assimilation of both observation and climate model data. CaPA outputs are highly valuable to modelling efforts dependent on precipitation inputs, and as such the quality of CaPA outputs is crucial. Two improvements to CaPA were investigated: reducing transformation bias though correction against moving-window averaged CaPA output that avoids transformation, and enhancing semivariograms through anisotropy and convection considerations. Accounting for convection in the semivariogram proved ineffectual, while the bias correction technique and anisotropic semivariograms both reduced bias and improved related metrics. No methods improved the Equitable Threat Score. If implemented separately, the bias correction or anisotropic semivariogram approaches will yield targeted benefits for CaPA users, particularly for applications focused on extreme precipitation values. Improvements were not so comprehensive as to warrant adoption in the operational CaPA configuration, although availability in experimental versions is recommended.

The Canadian Precipitation Analysis (CaPA) produces a gridded product by assimilating data from stations and the Global Environmental Multiscale (GEM) model. This project assesses the performance of the satellite based rainfall estimates for Canada, and the results of their assimilation with CaPA. The satellite based estimates considered are those from the Climate Prediction Center Morphing method (CMORPH) and Precipitation Estimation from Remotely Sensed Information using Artificial Neural Networks (PERSIANN). Relative to the Second Generation of Daily Adjusted Precipitation for Canada (APC2), all satellite products are shown to generally underestimate rainfall, however convective events result in an overestimation. Skill scores show that the satellite products possess the most skill for eastern Canada and decreasingly so westward. When assimilated with CaPA, the satellite products express decreased skill for light rainfall and potential improvements for larger events. While central Canada experiences the greatest improvements, all regions benefit the most from June through August.

No description available.

No description available.

No description available.

The intensity returns obtained by a radar from precipitation are well known to fluctuate violently in space and time. We present a systematic study of the resolution dependence time series with overlapping time resolutions spanning 10 orders of magnitude (0.77 ms to 4 months), of the fluctuating radar echo from precipitation. The results undermine the current assumptions of homogeneity of rainfield at scales smaller than the radar resolution, due to Marshall and Hitschfeld (1953), by showing that the only length scales identifiable in the time series are those of the radar pulse volume, the wavelength, and a very small inner scale of the order of millimeters. An analysis of the multiscaling nature of the time series of echo fluctuations reveals multiscaling behaviour at scales down to the resolution or pulse volume scale. Since there are no a priori scales in the rainfield we proceed to model the fluctuating radar echo by assuming a multiscaling model of rainfield variability which extends to sub-resolution scales. A systematic analysis of the statistical behaviour of computed reflectivities from this variability gives a full statistical description of reflectivity originating from multiscaling variability, and solves the scalar multifractal radar observer's problem. Computation of time series of reflectivities from a time-space representation of this variability reveals quantitative and qualitative behaviours consistent with those of observed echo fluctuation time series. We conclude that a multiscaling model of the rainfield which extends to the smallest scales of the rainfield is consistent with observation.