Premium
Habitat Suitability Criteria via Parametric Distributions: Estimation, Model Selection and Uncertainty
Author(s) -
Som Nicholas A.,
Goodman Damon H.,
Perry Russell W.,
Hardy Thomas B.
Publication year - 2016
Publication title -
river research and applications
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.679
H-Index - 94
eISSN - 1535-1467
pISSN - 1535-1459
DOI - 10.1002/rra.2900
Subject(s) - kernel density estimation , probability density function , density estimation , univariate , selection (genetic algorithm) , computer science , representation (politics) , parametric statistics , statistics , mathematics , multivariate statistics , machine learning , estimator , politics , political science , law
Previous methods for constructing univariate habitat suitability criteria (HSC) curves have ranged from professional judgement to kernel‐smoothed density functions or combinations thereof. We present a new method of generating HSC curves that applies probability density functions as the mathematical representation of the curves. Compared with previous approaches, benefits of our method include (1) estimation of probability density function parameters directly from raw data, (2) quantitative methods for selecting among several candidate probability density functions, and (3) concise methods for expressing estimation uncertainty in the HSC curves. We demonstrate our method with a thorough example using data collected on the depth of water used by juvenile Chinook salmon ( Oncorhynchus tschawytscha ) in the Klamath River of northern California and southern Oregon. All R code needed to implement our example is provided in the appendix. Published 2015. This article is a U.S. Government work and is in the public domain in the USA.