detecting hydrologic responses to earthquakes using support vector machines max rudolph

16
Detecting Hydrologic Responses to Earthquakes using Support Vector Machines Max Rudolph

Upload: justin-douglas

Post on 30-Dec-2015

217 views

Category:

Documents


0 download

TRANSCRIPT

Page 1: Detecting Hydrologic Responses to Earthquakes using Support Vector Machines Max Rudolph

Detecting Hydrologic Responses to Earthquakes using Support Vector Machines

Max Rudolph

Page 2: Detecting Hydrologic Responses to Earthquakes using Support Vector Machines Max Rudolph

Objectives/Questions• Why does stream flow respond or not respond to

earthquakes?– Effects of M, r, E/V, permeability/rock type, consolidated

vs. unconsolidated sediments?• Can we train a computer to automatically detect

these responses?– If so, we can mine a tremendous amount of stream flow

data.

Page 3: Detecting Hydrologic Responses to Earthquakes using Support Vector Machines Max Rudolph
Page 4: Detecting Hydrologic Responses to Earthquakes using Support Vector Machines Max Rudolph
Page 5: Detecting Hydrologic Responses to Earthquakes using Support Vector Machines Max Rudolph
Page 6: Detecting Hydrologic Responses to Earthquakes using Support Vector Machines Max Rudolph

BackgroundWhat is a Support Vector Machine?

It accepts a vector as input and returns a natural number corresponding to the class that the vector belongs to. The vector is called the ‘Feature Vector’. The SVM is finding a hyperplane that optimally divides parameter space into regions corresponding to each class.

Training:Provide SVM with paired set of feature vectors, labels. In this study, classes are ‘Yes’,

‘No’, ‘Maybe’ – determine 2 free parameters (C, gamma)

‘Linear’ kernel Radial Basis Function

(from

libS

VM w

ebsi

te)

Page 7: Detecting Hydrologic Responses to Earthquakes using Support Vector Machines Max Rudolph
Page 8: Detecting Hydrologic Responses to Earthquakes using Support Vector Machines Max Rudolph

Data SourcesData Sources:

USGS Instantaneous Data Archive (IDA)-79 stream gages available (with no upstream

regulation). 1-15 minute sample interval. 1980-2008NOAA National Climatic Data Center (NCDC) (230 stations)

NCEDC Earthquake Catalog 1980 to present Unexploited Data Sources (future work?):

NOAA hourly precipitation records 2001-presentUSGS daily stream flow archive - better availability

Page 9: Detecting Hydrologic Responses to Earthquakes using Support Vector Machines Max Rudolph

Data acquisition (done in Matlab!)myStation = station_numbers{i};

myURL = ['http://wdr.water.usgs.gov/wy2009/pdfs/' myStation '.2009.pdf']; myCMD = ['curl -O ' myURL];

conversionError = unix(['/sw/bin/pdftotext ' myStation '.2009.pdf']);%% check text files for phrase "No regulation or diversion upstream from station"noDiversions(i) = ~unix(['grep "No regulation or diversion upstream from station" ' txtFile]);if(noDiversions(i) == 0)

noDiversions(i) = ~unix(['grep "No regulation upstream from station" ' txtFile]);end

unix(['curl -c cookies.txt http://ida.water.usgs.gov/ida/available_records.cfm?sn=' siteid] unix(['curl -O -S -L -v -b cookies.txt -c cookies.txt -e http://ida.water.usgs.gov/ida/available_records.cfm?sn=' siteid ' -d "fromdate=' sds '&todate=' eds '&mindatetime=1980-01-04%2000%3A00%3A00.0&maxdatetime=2008-09-30%2023%3A45%3A00.0&site_no=' siteid '&rtype=1&submit1=Retrieve Data" http://ida.water.usgs.gov/ida/available_records_process.cfm’] unix(['mv available_records_process.cfm ' siteid '.rdb'])

Page 10: Detecting Hydrologic Responses to Earthquakes using Support Vector Machines Max Rudolph

Feature Vector1. Earthquake Magnitude/102. Log10(Earthquake Distance)/10

3. Normalized relative daily streamflow for each day in 7 days preceding-following earthquake

4. Histograms of normalized discharge in week after, before earthquake

5. Same as (4), but for first and second derivatives6. Same as (4) + (6), for smoothed stream data with 1 day

centered moving average filter7. Number of days with precipitation (any amount) at

nearest station in week preceding earthquake, week following earthquake

Page 11: Detecting Hydrologic Responses to Earthquakes using Support Vector Machines Max Rudolph

ResultsAfter training, cross validation indicates that the SVM

classifies the training dataset (500 samples) with 97% accuracy for (c, gamma) ~= (0.0016,0.3969).

When run on the full dataset for earthquakes with M>=3.5, of 113,365 feature vectors, 1121 are classified as ‘Yes’ or ‘Maybe’ (examples follow)

Main problem/shortcoming appears to be poor quality of climatic data.

Page 12: Detecting Hydrologic Responses to Earthquakes using Support Vector Machines Max Rudolph
Page 13: Detecting Hydrologic Responses to Earthquakes using Support Vector Machines Max Rudolph
Page 14: Detecting Hydrologic Responses to Earthquakes using Support Vector Machines Max Rudolph
Page 15: Detecting Hydrologic Responses to Earthquakes using Support Vector Machines Max Rudolph
Page 16: Detecting Hydrologic Responses to Earthquakes using Support Vector Machines Max Rudolph

Future directions1. Method could also be applied to well level data.

Wells vastly outnumber unregulated gaged streams.

2. Compare with Gleeson 2011 GRL (Permeability Map) – do rivers fed by low-k formations respond more or less readily than high-k formations?