From netCDF to GIS

1
From netCDF to GIS

• David R. Maidment
• Unidata Program Center, Boulder CO
• 6 Feb 2004

2
NetCDF and GIS

• NetCDF describes atmospheric and water properties
  varying continuously in (x,y,z,t)
• GIS describes the physical landscape,
• Vector GIS has spatially discrete features
  (point, line area)
• Raster GIS is spatially continuous (in 2D)
• How do we connect the water with the water
  environment

3
NetCDF File for Model Output
dimensions lat 5, lon 10, level 4, time
unlimited variables float rh (time, lat,
lon) int lat(lat), lon(lon) latunits
degrees_north longunits
degrees_east short time(time)
timeunits hours since 1996-1-1 data lat
20, 30, 40, 50, 60 long -160, -140, -118,
-96, -84, -52, -45, -35, -25, -15 time
12 rh .5,.2,.4,.2,.3,.2,.4,.5,.6,.7,
.1,.3,.1,.1,.1.,.1,.5,.7,.8,.8,
.1,.2,.2,.2,.2,.5,.7,.8,.9,.9,
.1,.2,.3,.3,.3,.3,.7,.8,.9,.9
.0,.1,.2,.4,.4,.4,.4,.7,.8,.9
4
Relative Humidity Points
5
Interpolate to Raster
GeoTiff format, cell size 0.5º
6
Zoom in to the United States
7
Average Rh in each State
Determined using Spatial Analyst function Zonal
Statistics with Rh as underlying raster and
States as zones
8
Unidata Toolbox
Relative Humidity Analysis Model
Created in Arc 9 Model Builder
9
Model Output Geodatabase
(Grids)
Geodatabase
Feature Dataset
Feature class
Feature class
Raster Catalog
Relationship
Raster
Table
Table
10
Raster Catalog has Types of Data
11
Scenarios for Data Dissemination

• ArcSDE and ArcIMS
• Leveraging the LDM
• Access through VisAD
• Direct downloads from ArcGIS

12
Scenario 1 ArcIMS ArcSDE

• Underlying data are stored in a relational
  database (e.g. Oracle)
• Accessed through Arc Spatial Data Engine (ArcSDE)
• Web access through Arc Internet Map Server
  (ArcIMS)
• Used for National Elevation Dataset, National
  Land Cover Dataset, National Hydrography Dataset
• Suitable for large static databases
• Possibly a good approach for historical climate
  datasets, e.g. VEMAP (NCAR Nan Rosenbloom, Dave
  Schimel, Tim Kittel)

National Elevation Dataset
13
Scenario 1 ArcIMS ArcSDE

• User Experience
• Go to a web site, browse available data and
  select a geographic region and desired data
  themes
• Download means that a data product is created
  off-line and user is notified by email when it is
  ready
• If lt 100MB download via ftp for free
• If gt 100MB write to CD-ROM and mail ()
• Contacts Paul Wiese, USGS, Denver (NHD) Sue
  Greenlee, USGS, EROS Data Center (NED, NLCD)

National Elevation Dataset
14
Scenario 2 Leverage the LDM

• File is disseminated through LDM network e.g.
  netCDF format
• Unidata publishes an Arc Toolbox for converting
  file (or adds a decoder to LDM to convert to
  Grid)
• Good for GIS analysis of existing real time data
  stream

15
Scenario 3 Access through VisAD

• VisAD is used to browse remote datasets
• Protocol to access Geography Network is added to
  VisAD
• Geographic data added to VisAD displays
• Atmospheric data converted to GeoTIFF for GIS
  analysis
• Build Raster Catalog of atmospheric datasets

Data Publisher 1
Data Publisher 4
Data Publisher 2
Data Publisher 3
Meteorologix publishes atmospheric data on the
Geography Network
16
Scenario 4 Direct Downloads from ArcGIS

• Select region and time period
• Launch query to http// access
• Download data
• Import to geodatabase
• Run models

Nexrad data for July 30, 2001, Rosillo Creek
17
Nexrad for Rosillo Creek
18
GIS in Water Resources Consortium
Water Resources

• Bringing together these two communities by using
  a common geospatial data model

19
Arc Hydro Hydrography
20
Arc Hydro Hydrology
21
Arc Hydro Framework Data Model
22
Arc Hydro Framework For South Florida
Basins
Waterbody (NHD)
Hydro Points
National Hydrography Dataset, NHD
23
Arc Hydro Components
24
Connecting Arc Hydro and Hydrologic Models
Interfacedata models
GIS
HMS
Geo Database
Arc Hydro data model
RAS
WRAP
25
Nexrad Map to Flood Map in Arc 9 Model Builder
Flood map as output
Model for flood flow
HMS
Model for flood depth
Nexrad rainfall map as input
26
Time and Space in GIS
Time Series
Feature Series
t3
t2
Value
t1
Time
Raster Series
Attribute Series
Value
t3
t2
t1
t1
t2
t3
y
x
27
Time Series and Temporal Geoprocessing
DHI Time Series Manager
Time Series
Feature Series
t3
t2
Value
t1
Time
Raster Series
Attribute Series
Value
t3
t2
t1
y
x
ArcGIS Temporal Geoprocessing
28
Time series from gages in Kissimmee Flood Plain

• 21 gages measuring water surface elevation
• Data telemetered to central site using SCADA
  system
• Edited and compiled daily stage data stored in
  corporate time series database called dbHydro
• Time series downloaded from dbHydro and stored
  for all gages as Arc Hydro Attribute Series

29
Arc Hydro Attribute Series
Map features have a time varying attribute
HydroID 2906
Feature Class (HydroID)
Attribute Series Table (FeatureID)
30
Raster Series
Ponded Water Depth Kissimmee River June 1, 2003
31
Arc Hydro Attribute Series Object
TSDateTime
Feature Class (point, line, area)
TSValue
FeatureID
TSType
TSType Table
32
NetCDF Data Model
Time
Dimensions and Coordinates
Value
Space (x,y,z)
Variable
Attributes
33
Conclusions

• Good time to make progress
• Unidata THREDDS research done at Unidata and some
  experience accumulated
• Arc 9.0 is about to be released with Model
  Builder planning for Arc 9.1 about to begin
• Fairly sound understanding of how to develop time
  series in ArcGIS
• Lots of cool weather and climate data that we
  really, really would like to use more!!