Spatially varying temperature trends in a Central California Estuary

  • Ricardo T. Lemos
  • Bruno Sansó
  • Marc Los Huertos
Article

DOI: 10.1198/108571107X227603

Cite this article as:
Lemos, R.T., Sansó, B. & Los Huertos, M. JABES (2007) 12: 379. doi:10.1198/108571107X227603

Abstract

We consider monthly temperature data collected over a period of 16 years at 24 stations in the estuarine wetlands of the Elkhorn Slough watershed, located in the Monterey Bay area in Central California, USA. Our goal is to develop a statistical model in order to separate the seasonal cycle, short-term fluctuations, and long-term trends, while accounting for the spatial variability of these features. In the model, each station has a specific, time-invariant mixture of two seasonal patterns, to encompass the spatial variability of oceanic influence. Likewise, trends are modeled as local mixtures of two patterns, to include the spatial variability of long-term temperature change. Finally, all stations share a common baseline, whose temporal variability is linearly dependent on a variable that summarizes several atmospheric measurements. We use a Bayesian approach with a purposely developed Markov chain Monte Carlo method to explore the posterior distribution of the parameters. We find that the seasonal cycles have changed in time, that neighboring stations can have substantially different behaviors, and that most stations show significant warming trends.

Key Words

Bayesian modeling Mixture models Space-time models Time-varying trends Water quality data 

Copyright information

© International Biometric Society 2007

Authors and Affiliations

  • Ricardo T. Lemos
    • 1
  • Bruno Sansó
    • 2
  • Marc Los Huertos
    • 3
  1. 1.Institute of OceanographyUniversidade de Lisboa and Maretec-Instituto Superior Técnico, Universidade Técnica de LisboaPortugal
  2. 2.Department of Applied Mathematics and StatisticsUniversity of CaliforniaSanta Cruz
  3. 3.Science and Environmental PolicyCalifornia State University Monterey BaySeaside

Personalised recommendations