In this paper, a statistical inference of Southeastern Canada extreme daily precipitation amounts is proposed using a classical nonstationary peaks-over-threshold model. Indeed, the generalized Pareto distribution (GPD) is fitted to excess time series derived from annual averages of independent precipitation amount events above a fixed threshold, the 99th percentile. Only the scale parameter of the fitted distribution is allowed to vary as a function of a covariate. This variability is modeled using B-spline function. Nonlinear correlation and cross-wavelet analysis allowed identifying two dominant climate indices as covariates in the study area, Arctic Oscillation (AO) and Pacific North American (PNA). The nonstationary frequency analysis showed that there is an east-west behavior of the AO index effects on extreme daily precipitation amounts in the study area. Indeed, the higher quantiles of these events are conditional to the AO positive phase in Atlantic Canada, while those in the more southeastern part of Canada, especially in Southern Quebec and Ontario, are negatively related to AO. The negative phase of PNA also gives the best significant correlation in these regions. Moreover, a regression analysis between AO (PNA) index and conditional quantiles provided slope values for the positive phase of the index on the one hand and the negative phase and on the other hand. This statistic allows computing a slope ratio which permits to sustain the nonlinear relation assumption between climate indices and precipitation and the development of the nonstationary GPD model for Southeastern Canada extremes precipitation modeling.
Quantile estimates are generally interpreted in association with the return period concept in practical engineering. To do so with the peaks‐over‐threshold (POT) approach, combined Poisson‐generalized Pareto distributions (referred to as PD‐GPD model) must be considered. In this article, we evaluate the incorporation of non‐stationarity in the generalized Pareto distribution (GPD) and the Poisson distribution (PD) using, respectively, the smoothing‐based B‐spline functions and the logarithmic link function. Two models are proposed, a stationary PD combined to a non‐stationary GPD (referred to as PD0‐GPD1) and a combined non‐stationary PD and GPD (referred to as PD1‐GPD1). The teleconnections between hydro‐climatological variables and a number of large‐scale climate patterns allow using these climate indices as covariates in the development of non‐stationary extreme value models. The case study is made with daily precipitation amount time series from southeastern Canada and two climatic covariates, the Arctic Oscillation (AO) and the Pacific North American (PNA) indices. A comparison of PD0‐GPD1 and PD1‐GPD1 models showed that the incorporation of non‐stationarity in both POT models instead of solely in the GPD has an effect on the estimated quantiles. The use of the B‐spline function as link function between the GPD parameters and the considered climatic covariates provided flexible non‐stationary PD‐GPD models. Indeed, linear and nonlinear conditional quantiles are observed at various stations in the case study, opening an interesting perspective for further research on the physical mechanism behind these simple and complex interactions.
Climate impact studies often require a reduction of the ensembles of opportunity from the Coupled Model Intercomparison Project when the simulations are used to drive impact models. An impact model’s nature limits the number of feasible realizations based on complexity and computational requirements or capacities. For the purpose of driving a hydrological model and an ocean model in the BaySys research program, two hierarchical, differently sized simulation ensembles were produced to represent climate evolution for the region of the Hudson Bay Drainage Basin. We compare a 19-member ensemble to a 5-member subset to demonstrate comparability of the driving climate used to produce model results. Ten extreme climate indicators and their changes are compared for the full study region and seven sub regions, on an annual and seasonal basis and for two future climate horizons. Results indicate stronger warming in the North and for cold temperatures and an East-West gradient in precipitation with larger absolute increases to the East and South of the Hudson Bay. Generally, the smaller ensemble is sufficient to adequately reproduce the mean and spread in the indicators found for the larger ensemble. The analysis of extreme climate indicators ensures that the tails of the distribution of temperature and precipitation are addressed. We conclude that joint analysis at the interface of the hydrological and ocean model domains are not limited by the application of differently sized climate simulation ensembles as driving input for the two different modeling exercises of the BaySys project environmental studies, yet acknowledging that impact model output may be dependent on other factors.
The availability of hydrometric data, as well as its spatial distribution, is important for water resources management. An overly dense network or an under developed network can cause inaccurate hydrological regional estimates. The objective of this study is to propose a methodology for rationalizing a network, specifically the New Brunswick Hydrometric Network. A hierarchical clustering analysis allowed dividing the province into two regions (North and South), based on latitude and high flow timing. These groups were subsequently split separately into three homogeneous subgroups, based on the generalized extreme value (GEV) distribution shape parameter of each station for annual maximum flow series. An entropy method was then applied to compute the amount of information shared between stations, ranking each station’s importance. A station with a lot of shared information is redundant (less important), whereas one with little shared information is unique (very important). The entropy method appears to be a useful decisional tool in a network rationalization.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.
hi@scite.ai
10624 S. Eastern Ave., Ste. A-614
Henderson, NV 89052, USA
Copyright © 2024 scite LLC. All rights reserved.
Made with 💙 for researchers
Part of the Research Solutions Family.