東華大學圖書館 |

Language: English

Help

回圖書館首頁

手機版館藏查詢

Back

Switch To: Labeled | MARC Mode | ISBD

Linked to FindBook

Google Book

Amazon

博客來

Predicting Flow Duration and Assessing Its Drivers in North-Central Colorado Using Crowdsourced Data.

Record Type:	Electronic resources : Monograph/item
Title/Author:	Predicting Flow Duration and Assessing Its Drivers in North-Central Colorado Using Crowdsourced Data./
Author:	Peterson, David.
Published:	Ann Arbor : ProQuest Dissertations & Theses, : 2022,
Description:	49 p.
Notes:	Source: Masters Abstracts International, Volume: 83-12.
Contained By:	Masters Abstracts International83-12.
Subject:	Water resources management. -
Online resource:	http://pqdd.sinica.edu.tw/twdaoapp/servlet/advanced?query=28967371
ISBN:	9798438793175

Predicting Flow Duration and Assessing Its Drivers in North-Central Colorado Using Crowdsourced Data.
Peterson, David.

Predicting Flow Duration and Assessing Its Drivers in North-Central Colorado Using Crowdsourced Data. - Ann Arbor : ProQuest Dissertations & Theses, 2022 - 49 p.

Source: Masters Abstracts International, Volume: 83-12.

Thesis (M.S.)--Colorado State University, 2022.

This item must not be sold to any third party vendors.

Headwater streams are globally important both ecologically and for human resource needs. These streams represent the majority of stream network length, but their flow regimes are often unknown. Streams can be classified by flow regime as perennial, intermittent, or ephemeral. These classifications are used in forest land management decisions and may affect Clean Water Act jurisdiction; however, the National Hydrography Dataset (NHD) often misclassifies headwater streams. The goal of this study is to model flow duration across the stream networks of eight subbasins in north-central Colorado. We used crowdsourced flow presence/absence data from 82 sites in the Stream Tracker program and eight flow sensors to train random forest regression models; these models predicted the fraction of time a stream flows from April-September for both the average from 2016-2020 (dubbed mean annual) and yearly averages (annual). Model predictor variables included climatic, physiographic, and land cover attributes of the study area. Models were developed using a sample of the sites for training and leaving the remaining sites for model testing. The resulting mean annual model's Nash-Sutcliffe efficiency (NSE) was 0.88 for test data, and the annual model's test data had an NSE value of 0.81. We found climate variables such as snow persistence, precipitation, and potential evapotranspiration most influential in predicting flow fraction based on the random forest-ranked variable importance. Forested and herbaceous land cover as well as depth to bedrock, available water storage, hydraulic conductivity, hydrologic soil group, drainage area, and watershed curvature were also identified as important drivers. We developed maps of predicted flow fractions and compared them to NHD flow classifications. In the Cache La Poudre subbasin, the mean annual model predicted perennial flow in 10% of streams and intermittent or ephemeral flow in 90% of streams. Our model predicted nonperennial flow for 76% of the streams that were mapped as perennial in the medium-resolution NHD. Based on these findings, the NHD over-represented perennial streams, classifying them three times more than our model, and under-represented intermittent and ephemeral streams by 32% in our study area. The annual model captured interannual variability in flow fraction and highlighted isolated areas of high variability in flow fraction between years in mid-to-low elevations. The models we developed using crowdsourced data can improve flow classifications of headwater streams and inform resource management decisions in northern Colorado. Crowdsourced streamflow data can be used in streamflow predictions anywhere that nonperennial flow is common.

ISBN: 9798438793175Subjects--Topical Terms:

794747
Water resources management.
Subjects--Index Terms:

Crowdsourcing

Predicting Flow Duration and Assessing Its Drivers in North-Central Colorado Using Crowdsourced Data.
LDR:03874nmm a2200361 4500 001 2348198
005 20220908122958.5
008 241004s2022 ||||||||||||||||| ||eng d
020 $a 9798438793175
035 $a (MiAaPQ)AAI28967371
035 $a AAI28967371
040 $a MiAaPQ $c MiAaPQ
100 1 $a Peterson, David. $3 3687518
245 1 0 $a Predicting Flow Duration and Assessing Its Drivers in North-Central Colorado Using Crowdsourced Data.
260 1 $a Ann Arbor : $b ProQuest Dissertations & Theses, $c 2022
300 $a 49 p.
500 $a Source: Masters Abstracts International, Volume: 83-12.
500 $a Advisor: Kampf, Stephanie K.
502 $a Thesis (M.S.)--Colorado State University, 2022.
506 $a This item must not be sold to any third party vendors.
520 $a Headwater streams are globally important both ecologically and for human resource needs. These streams represent the majority of stream network length, but their flow regimes are often unknown. Streams can be classified by flow regime as perennial, intermittent, or ephemeral. These classifications are used in forest land management decisions and may affect Clean Water Act jurisdiction; however, the National Hydrography Dataset (NHD) often misclassifies headwater streams. The goal of this study is to model flow duration across the stream networks of eight subbasins in north-central Colorado. We used crowdsourced flow presence/absence data from 82 sites in the Stream Tracker program and eight flow sensors to train random forest regression models; these models predicted the fraction of time a stream flows from April-September for both the average from 2016-2020 (dubbed mean annual) and yearly averages (annual). Model predictor variables included climatic, physiographic, and land cover attributes of the study area. Models were developed using a sample of the sites for training and leaving the remaining sites for model testing. The resulting mean annual model's Nash-Sutcliffe efficiency (NSE) was 0.88 for test data, and the annual model's test data had an NSE value of 0.81. We found climate variables such as snow persistence, precipitation, and potential evapotranspiration most influential in predicting flow fraction based on the random forest-ranked variable importance. Forested and herbaceous land cover as well as depth to bedrock, available water storage, hydraulic conductivity, hydrologic soil group, drainage area, and watershed curvature were also identified as important drivers. We developed maps of predicted flow fractions and compared them to NHD flow classifications. In the Cache La Poudre subbasin, the mean annual model predicted perennial flow in 10% of streams and intermittent or ephemeral flow in 90% of streams. Our model predicted nonperennial flow for 76% of the streams that were mapped as perennial in the medium-resolution NHD. Based on these findings, the NHD over-represented perennial streams, classifying them three times more than our model, and under-represented intermittent and ephemeral streams by 32% in our study area. The annual model captured interannual variability in flow fraction and highlighted isolated areas of high variability in flow fraction between years in mid-to-low elevations. The models we developed using crowdsourced data can improve flow classifications of headwater streams and inform resource management decisions in northern Colorado. Crowdsourced streamflow data can be used in streamflow predictions anywhere that nonperennial flow is common.
590 $a School code: 0053.
650 4 $a Water resources management. $3 794747
653 $a Crowdsourcing
653 $a Flow duration
653 $a Intermittent streams
653 $a Nonperennial streams
653 $a Random forest
653 $a Colorado
690 $a 0595
710 2 $a Colorado State University. $b Ecosystem Science and Sustainability. $3 3546004
773 0 $t Masters Abstracts International $g 83-12.
790 $a 0053
791 $a M.S.
792 $a 2022
793 $a English
856 4 0 $u http://pqdd.sinica.edu.tw/twdaoapp/servlet/advanced?query=28967371