Search Results

You are looking at 1 - 3 of 3 items for

  • Author or Editor: P. Guttorp x
  • Refine by Access: All Content x
Clear All Modify Search
James P. Hughes
and
Peter Guttorp

Abstract

Nonhomogeneous hidden Markov models (NHMM) provide a method of relating synoptic atmospheric measurements to precipitation occurrence at a network of rain gauge stations. In previous work it was assumed that, conditional on the current atmospheric pattern (termed a “weather state”), rain gauge stations in a network could be considered spatially independent. For a spatially dense network, this assumption is not tenable. In the present work, the NHMM is extended to include the case of spatial dependence by postulating an autologistic model for the conditional probability of rainfall given the weather state. Methods for fitting the parameters, assessing the goodness of fit of the model, and generating rainfall simulations are presented. The model is applied to a network of 24 stations in the Puget Sound region of western Washington State.

Full access
M. L. Thompson
and
P. Guttorp

Abstract

We consider data on severe cyclonic storms striking the Day of Bengal coast during the period 1877–1977. In the literature these data have been modeled by a homogeneous Poisson process in which case times between storm occurrences are independent of one another, making prediction, and hence advance planning, impossible. We give some evidence against the adequacy of a Poisson process model and suggest a Poisson cluster model that appears to describe the data better. The features of fills model are such as to enable some planning procedures to he developed.

Full access
Le Bao
,
Tilmann Gneiting
,
Eric P. Grimit
,
Peter Guttorp
, and
Adrian E. Raftery

Abstract

Wind direction is an angular variable, as opposed to weather quantities such as temperature, quantitative precipitation, or wind speed, which are linear variables. Consequently, traditional model output statistics and ensemble postprocessing methods become ineffective, or do not apply at all. This paper proposes an effective bias correction technique for wind direction forecasts from numerical weather prediction models, which is based on a state-of-the-art circular–circular regression approach. To calibrate forecast ensembles, a Bayesian model averaging scheme for directional variables is introduced, where the component distributions are von Mises densities centered at the individually bias-corrected ensemble member forecasts. These techniques are applied to 48-h forecasts of surface wind direction over the Pacific Northwest, using the University of Washington mesoscale ensemble, where they yield consistent improvements in forecast performance.

Full access