Letteratura scientifica selezionata sul tema "Multisite stochastic downscaling"

Abstract A Bernoulli–generalized Pareto multivariate autoregressive (BMAR) model is proposed in this paper for multisite statistical downscaling of daily precipitation. The proposed model relies on a probabilistic framework to describe the conditional probability density function of precipitation at each station for a given day and handles multivariate dependence in both time and space using a multivariate autoregressive model. Within a probabilistic framework, BMAR employs a regression model whose outputs are parameters of the mixed Bernoulli–generalized Pareto distribution. As a stochastic component, the BMAR employs a latent multivariate autoregressive Gaussian field to preserve lag-0 and lag-1 cross correlations of precipitation at multiple sites. The proposed model is applied for the downscaling of AOGCM data to daily precipitation in the southern part of Québec, Canada. Reanalysis products are used in this study to assess the potential of the proposed method. Based on the mean errors (MEs), the root-mean-square errors (RMSEs), precipitation indices, and the ability to preserve lag-0 and lag-1 cross correlation, results of the study indicate the superiority of the proposed model over a multivariate multiple linear regression (MMLR) model and a multisite hybrid statistical downscaling procedure that combines MMLR and stochastic generator schemes.

Abstract A nonlinear, probabilistic synoptic downscaling algorithm for daily precipitation series at multiple sites is presented. The expanded Bernoulli–gamma density network (EBDN) represents the conditional density of multisite precipitation, conditioned on synoptic-scale climate predictors, using an artificial neural network (ANN) whose outputs are parameters of the Bernoulli–gamma distribution. Following the methodology used in expanded downscaling, predicted covariances between sites are forced to match observed covariances through the addition of a constraint to the ANN cost function. The resulting model can be thought of as a regression-based downscaling model with a stochastic weather generator component. Parameters of the Bernoulli–gamma distribution are downscaled from the synoptic-scale circulation, and unresolved temporal variability is generated via an autoregressive noise model. Demonstrated on a multisite precipitation dataset from coastal British Columbia, Canada, the EBDN is capable of specifying the conditional distribution of precipitation at each site, modeling the occurrence and the amount of precipitation simultaneously, reproducing observed spatial relationships between sites, randomly generating realistic synthetic precipitation series, and predicting precipitation amounts in excess of those in the observational record.

This study presents an exhaustive evaluation of the performance of three statistical downscaling techniques for generating daily rainfall occurrences at 22 rainfall stations in the upper Ping river basin (UPRB), Thailand. The three downscaling techniques considered are the modified Markov model (MMM), a stochastic model, and two variants of regression models, statistical models, one with single relationship for all days of the year (RegressionYrly) and the other with individual relationships for each of the 366 days (Regression366). A stepwise regression is applied to identify the significant atmospheric (ATM) variables to be used as predictors in the downscaling models. Aggregated wetness state indicators (WIs), representing the recent past wetness state for the previous 30, 90 or 365 days, are also considered as additional potential predictors since they have been effectively used to represent the low-frequency variability in the downscaled sequences. Grouping of ATM and all possible combinations of WI is used to form eight predictor sets comprising ATM, ATM-WI30, ATM-WI90, ATM-WI365, ATM-WI30&90, ATM-WI30&365, ATM-WI90&365 and ATM-WI30&90&365. These eight predictor sets were used to run the three downscaling techniques to create 24 combination cases. These cases were first applied at each station individually (single site simulation) and thereafter collectively at all sites (multisite simulations) following multisite downscaling models leading to 48 combination cases in total that were run and evaluated. The downscaling models were calibrated using atmospheric variables from the National Centers for Environmental Prediction (NCEP) reanalysis database and validated using representative General Circulation Models (GCM) data. Identification of meaningful predictors to be used in downscaling, calibration and setting up of downscaling models, running all 48 possible predictor combinations and a thorough evaluation of results required considerable efforts and knowledge of the research area. The validation results show that the use of WIs remarkably improves the accuracy of downscaling models in terms of simulation of standard deviations of annual, monthly and seasonal wet days. By comparing the overall performance of the three downscaling techniques keeping common sets of predictors, MMM provides the best results of the simulated wet and dry spells as well as the standard deviation of monthly, seasonal and annual wet days. These findings are consistent across both single site and multisite simulations. Overall, the MMM multisite model with ATM and wetness indicators provides the best results. Upon evaluating the combinations of ATM and sets of wetness indicators, ATM-WI30&90 and ATM-WI30&365 were found to perform well during calibration in reproducing the overall rainfall occurrence statistics while ATM-WI30&365 was found to significantly improve the accuracy of monthly wet spells over the region. However, these models perform poorly during validation at annual time scale. The use of multi-dimension bias correction approaches is recommended for future research.

This article reviews the historical development of statistical weather models, from simple analyses of runs of consecutive rainy and dry days at single sites, through to multisite models of daily precipitation. Weather generators have been used extensively in water engineering design and in agricultural, ecosystem and hydrological impact studies as a means of in-filling missing data or for producing indefinitely long synthetic weather series from finite station records. We begin by describing the statistical properties of the rainfall occurrence and amount processes which are necessary precursors to the simulation of other (dependent) meteorological variables. The relationship between these daily weather models and lower-frequency variations in climate statistics is considered next, noting that conventional weather generator techniques often fail to capture wholly interannual variability. Possible solutions to this deficiency - such as the use of mixtures of slowly and rapidly varying conditioning variables - are discussed. Common applications of weather generators are then described. These include the modelling of climate-sensitive systems, the simulation of missing weather data and statistical downscaling of regional climate change scenarios. Finally, we conclude by considering ongoing advances in the simulation of spatially correlated weather series at multiple sites, the downscaling of interannual climate variability and the scope for using nonparametric techniques to synthesize weather series.

This thesis presents the development and application of a downscaling framework for multi site simulation of daily rainfall. The rainfall simulation is achieved in two stages. First, rainfall occurrences at multiple sites are downscaled, which is followed by the generation of daily rainfall amounts at each site identified as wet. A continuous weather state based nonparametric downscaling model conditional on atmospheric predictors and a previous day average rainfall state is developed for simulation of multi site rainfall occurrences. A nonparametric kernel density approach is used for simulation of rainfall amounts at individual sites conditional on atmospheric variables and the previous day rainfall amount. The proposed model maintains spatial correlation of rainfall occurrences by simulating concurrently at all stations and of amounts by using random innovations that are spatially correlated yet serially independent. Temporal dependence is reproduced in the occurrence series by conditioning on previous day average wetness fraction and assuming the weather states to be Markovian, and in the amount series by conditioning on the previous day rainfall amount. The seasonal transition is maintained by simulating rainfall on a day-to-day basis using a moving window formulation. The developed downscaling framework is calibrated using the relevant atmospheric variables and rainfall records of 30 stations around Sydney, Australia. Results indicate a better representation of the spatio-temporal structure of the observed rainfall as compared to existing alternatives. Subsequently, the framework is applied to predict plausible changes in rainfall in warmer conditions using the same set of atmospheric variables for future climate obtained as a General Circulation Model simulation. While the case studies presented are restricted to a specific region, the downscaling model is designed to be useful in any generic catchment modelling and management activity and/or for investigating possible changes that might be experienced by hydrological, agricultural and ecological systems in future climates.