Visualization and Analysis of Distributed Hydrologic Model Results Using Open-source
Software Tools
3rd CUAHSI Conference on Hydroinformatics July 15-17, 2015
Thomas Adams [email protected]
Outline Significant features of open source tools for
Hydroinformatics
NEXRAD-based Stage3/MPE bias analysis and correction RDHM Distributed Hydrologic Modeling parameter
estimation & simulation visualization
Flood inundation mapping; probabilistic flood inundation mapping
Geostatistical simulations random field generation Publication quality graphics: GRASS GIS, R, GMT
Open Source tools for Hydroinformatics
GRASS GIS, QGIS, SAGA GIS, R, Generic Mapping Tools (GMT), Octave (~Matlab clone)
Linux, Mac OS X, MS-Windows PostgreSQL, SQLite3, MySQL,PostGIS, Mapserver, GeoServer OSM-WMS support Data types: raster, vector, image (satellite), 3D raster, raster timeseries Data formats: grib, grib2, netcdf, hdf5, shapefiles, ascii, Arc Ascii grid
virtually all formats supported through gdal
Fully extensible (add-on modules) and source code changes (C, FORTRAN, Python)
Easily scriptable with Linux shell scripting, Python, Perl Map projection and datum transformations Advanced GUIs
NLDAS vs PRISM precipitationDec 2007
Washington, DC
NWS NEXRADRadar Locations
NEXRAD based Stage3/MPE Precipitation3 May 2002, 24-hr Accumulation
Bias = MPE/PRISM
Distributed Hydrologic Model Parameter Estimation
ISRIC Global Soils Database SSURGO in the U.S.
From: Koren, et al NOAA/NWS/NWC
SAC-SMA Parameters
From: Koren, et al NOAA/NWS/NWC
Soil Hydraulic Properties from Soil Texture
From: Koren, et al NOAA/NWS/NWC
SAC-SMA LZTWM
Source: NOAA/NWS/NWC
Estimated SAC-SMA LZTWM
GRASS GIS 3D NVIZ VisualizationVillahermosa, Tabasco Mexico
Tabasco, MXRDHM Routing Parameter, alpha
250m
01 0
0 02 0
0 03 0
0 04 0
0 05 0
0 06 0
0 0
RDHM simulation vs USGS observed Potomac R at L Falls Pump Station
Oct 2007 Dec. 2013
Date
D is c
h ar g
e ( C
MS )
LegendRDHMUSGS
10/20/2009 11/20/2009 12/20/2009 01/20/2010 02/20/2010 03/20/2010 04/20/2010 05/20/2010
qm = 1.39
Uncalibrated RDHM SAC-SMA
05 0
1 00
1 50
2 00
RDHM simulation vs USGS observed Seneca Creek near Dawsonville, MD
Jan Dec 2003
Date
D is c
h ar g
e ( C
MS )
LegendSimulatedObserved
01/01/2003 03/01/2003 05/01/2003 07/01/2003 09/01/2003 11/01/2003
Bankfull Flow = 85 cms
Flow Exceeding Bankfull Flowbased on USGS Regional Analyses
Flow (cms)
RDHM sac visualization4km spatial resolution
Discharge Precipitation
Flows 1 cms, 250m resolution
Discharge Precipitation
Coupling RDHM sac with HEC-RAS unsteady flow simulations
MMEFS NAEFS Hydrologic Ensembles
http://www.erh.noaa.gov/mmefs/
Potomac River, Washington, DC
USGS 3m DEM
Google Earth Visualization from GRASS kmz file
Geostatistical Random Field Generation
GRASS with gstat
24-hour Precipitation ending 07/04/2006 12UTC
MM5 NEXRAD MPE
Randomly Sample Precipitation and Temperature Fields
-or-
Soil Moisture Field
4 Realizations of Conditional Gaussian Geostatistical Simulations using R/gstat
and GRASS GIS
Unconditional Gaussian Simulation
GRASS & Generic Mapping Tools (GMT) working together
From dylan beaudette USDA/NRCS
Draw your attention to
FLOODFORECASTING
A GLOBAL PERSPECTIVE
THOMAS E. ADAMS, IIITHOMAS C. PAGANO Late 2015/Early 2016
Chapters on National Systems in: Australia Brazil (Amazon Basin) China Columbia EU EFAS Germany India Israel Romania Russia Sweden (hopefully) UK USA also Congo Basin USGS chapter on Streamgauging
End