Graduate Project

 

An evaluation of temperature prediction models on Coastal Oregon and western Cascade streams Public Deposited

Downloadable Content

Download PDF
https://ir.library.oregonstate.edu/concern/graduate_projects/44558k38f

Descriptions

Attribute NameValues
Creator
Abstract
  • Evaluations were performed on three stream temperature prediction models: Heat Source 5.6, SSTEMP 3.9, and Brown's equation. Each model was evaluated using data collected from three Coastal Oregon and two western Cascade streams. Stream temperature simulations were performed on two forested reaches, one clearcut reach, and one buffered reach, for each site. One hot day, within five to seven consecutive hot days, was simulated. Each model was evaluated on sensitivity, accuracy, precision, and practicality. Sensitivity evaluations were performed to determine the input parameters that had the greatest effect on modeled temperatures and that helped determine the precision needed for input parameters. Sensitivity was quantified as the number of sensitive parameters compared to the total number of input parameters. The percentage of sensitive parameters for Heat Source, SSTEMP, and Brown's equation were 12, 30, and 100 percent, respectively. Model accuracy was evaluated to determine how well the predicted stream temperatures matched the observed stream temperatures. Precision was evaluated to determine consistency, bias, and the overall average error for all study sites combined. Heat Source demonstrated the highest model accuracy and precision. However, a double-peaked pattern in hourly stream temperature predictions showed deviations from the observed temperatures by as much as 3.5°C in buffered and some forested areas. This pattern was never observed in the stream temperature data and could indicate a potential flaw in the model. Heat Source also consistently predicted maximum daily stream temperatures prior to the observed maximum daily temperatures. Brown's equation demonstrated high model accuracy and precision with only three input variables. SSTEMP demonstrated relatively poor model accuracy and precision, especially in clearcut treatments. However, the model's strongest predictive ability is in daily mean stream temperature. S STEMP under-predicted 95 percent of the observed minimum temperatures and over-predicted 90 percent of the observed maximum temperatures. Evaluations of practicality considered the cost and user-friendliness of each model. All models are in the public domain and may be obtained for free by internet or mail. Collection of input data for Heat Source and SSTEMP require the most time and resources because of the large number of input parameters. Brown's equation does not require a field visit, however one will increase prediction accuracy. The input of data and the simulation run times were generally less than two minutes for all models. Heat Source is very user-friendly primarily because of graphic output and userscreens. Brown's equation is also user-friendly primarily because it is simple and requires no computer. SSTEMP was user-friendly in some aspects, yet outdated in others like the MS-DOS interface and absence of graphic output. Results in this document are applicable only to hot summer days. Use of the models for other climatic conditions will, most likely, work but were not validated by this project.
Resource Type
Date Available
Date Issued
Degree Level
Degree Name
Degree Field
Degree Grantor
Commencement Year
Advisor
Committee Member
Academic Affiliation
Rights Statement
Publisher
Peer Reviewed
Language
Replaces

Relationships

Parents:

This work has no parents.

Items