BOLIVIAN ALTIPLANO GIS DATASET

By Cesar Gomez Paez. (22/02/2012).

After the publication of 'An approach to the hydraulic culture on the Bolivian
Altiplano ' I've created a minimal GIS dataset for the Bolivian Altiplano in two common
formats: SHP (ESRI shapefile format) and KML (Keyhole Markup Language). This dataset
contains three diferent data types:

• Geological information derived for the data obtained from the USGS.

• Watershed data derived from the SRTM obtained from the GLCF.

• Custom data, vectorized by the author.

Figure 1. An overview of Bolivian Altiplano GIS dataset showed in Google Earth.

DATA PROCESING OF GEOLOGICAL INFORMATION.

The geodata obtained from the USGS comes in E00 file format (ARC/INFO interchange
file format), this type of file format is not well suited to work with it directly so I proceeded to
transform it in a more common file format, like the SHP file format. From USGS ... 'The
geologic map of the Altiplano/Cordillera Occidental was produced digitally from fifteen,
1:250,000 scale, field compilation maps. These paper maps were made by Servicio Geológico
de Bolivia (GEOBOL) geologists over the last decade'. If you want to download the original
E00 coverages go to DATA SOURCES section at the end of this paper, there you can access
the direct online links to the original sources of the data.
 The geodata comprise the next files:

FILE DESCRIPTION
altidis.new.e00 ARC/INFO export file of the geologic map
labels.e00 ARC/INFO export file of labels for geologic map
struct.all.e00 ARC/INFO export file of the structural geology map
volmajor.e00 ARC/INFO export file of major volcanoes
volminor.e00 ARC/INFO export file of minor volcanoes
Table 1. ARC/INFO export format Description of coverage.

This dataset lacks in the projection files so I proceeded to create the projection files
based on the information contained in the metadata text:

Spatial_Reference_Information:

Horizontal_Coordinate_System_Definition:

Planar:

Map_Projection:

Map_Projection_Name: Lambert Conformal Conic

Lambert_Conformal_Conic:

Standard_Parallel: -17.333333
Standard_Parallel: -22.666667
Longitude_of_Central_Meridian: -67.5
Latitude_of_Projection_Origin: -24
False_Easting: 0
False_Northing: 0

Geodetic_Model:

Horizontal_Datum_Name: D_International_1924
Ellipsoid_Name: International_1924
Semi-major_Axis: 6378388.0
Denominator_of_Flattening_Ratio: 297.0

The projection files are simple text files with a few lines, I show one of this files
(altidis.new.prj), the region covered is the same for all the E00 files:

Projection LAMBERT
Datum WGS84
Zunits NO
Units METERS
Xshift 0.000000
Yshift 0.000000
Parameters
-17.333333 /* 1st standard parallel
-22.666667 /* 2nd standard parallel
-67.5 /* central meridian
-24.0 /* latitude of the origin
0.000000 /* false easting (meters)
0.000000 /* false northing (meters)
 I can't deal with the D_International_1924 Datum so I substituted it by WGS84
Datum, the result of this substitution fits well in the real world.

Then I show a frame taken from within Quantum GIS (an OpenSource GIS software)
of the reprojected geological layer.

Figure 2. Geological layer, white polygons (EPSG:4326).

After the transformation, all geological layers come in SHP file format and have
Geographic Projection and WGS84 Datum (EPSG:4326).

FILE DESCRIPTION
altidis.new_polygon.shp ESRI shapefile of the geologic map.
altidis.new_point.shp ESRI shapefile of the geologic map.
struct.all_line.shp ESRI shapefile of the structural geology map.
volmajor_point.shp ESRI shapefile of major volcanoes
volminor_point.shp ESRI shapefile of minor volcanoes
Table 2. List of ESRI shapefiles. Description of coverage.

The original altidis.new.E00 comes with the next data fields:

AREA PERIMETER ALTIDIS.NE ALTIDIS.N1 GEO1 GRETA

The geologic info is coded in GEO1, a numeric field, of course this's not usefull so I
added a new text data field called DESCIPTION. I filled this new data field with the info
contained in metadata.txt. This text file contains the GEO1 numeric field as well the text with
the geological description of that code. Take in mind the restriction applied under the DBF files,
the string field only can hold 255 characters, due to this limitation in some codes not all the
text has been included. I show a table with the numeric codes and its text equivalence:
 Digital Geologic Map of the Altiplano and Cordillera Occidental (METADATA).

GEO1 DESCIPTION

1001 Bodies of water.

400 Gypsum diapirs (Miocene to Eocene?).

May include halite and other evaporite minerals. Includes deposits mapped as Chuquichambi
Formation and Jalluma Formation.

300 Sedimentary rocks (Paleozoic).

Chiefly marine sandstone and shale of Devonian to Ordovician age. Rocks are generally highly
folded and locally penetratively deformed. Includes deposits mapped as Amatura Formation,
Cancañiri Formation, Catavi Formation, Colpachuco Formation, Copacabana Formation, Cumana
Formation, Huanuni Formation, Khasa Formation, Llallagua Formation, Sica Sica Formation, Uncia
Formation, and Vila Vila Formation.

200 Sedimentary rocks (Cretaceous).

Marine and nonmarine sandstone, shale, marl, and limestone. Includes deposits mapped as
Aroifilla Formation, El Molino Formation, La Puerta Formation, Miraflores Formation, and Tarapaya
Formation.

160 Ks? Three polygons with this attribute value occur in the data. One is a sliver polygon, and the
other two are symbolized as Ks on the printed map.

150 Los Frailes and Morococala Ignimbrites (Miocene).

Ash-flow tuffs of dacitic composition in extensive ignimbrite field exposed primarily east of map
area.

140 Sedimentary rocks (Oligocene to Paleocene).

Nonmarine, mostly reddish colored conglomerate, sandstone, shale, and mudstone. Primary host
for sedimentary-hosted copper deposits. Includes deposits mapped as Berenguela Formation,
Nazareno Formation, Potoco Formation, Santa Lucia Formation, San Andres Formation, and
Tihuanacu Formation.

130 Sedimentary rocks (Pliocene to Oligocene).

Nonmarine sandstone, conglomerate, shale, marl, and evaporites. Includes deposits mapped as
Azurita Formation, Ballivian Formation, Coniri Formation, Cosuño Formation, Crucero Formation,
Huayllamarca Formation, Huayllapucara Formation, Kasira Formation, Lower Quehua Formation,
Mondragon Formation, Pizaqueri Formation, San Vicente Formation, Tambillo Formation, Totora
Formation, and Turco Formation.

120 Pyroclastic rocks (Miocene and Oligocene).

Chiefly welded to nonwelded ash-flow tuffs, but includes air-fall tuffs and thin, volcaniclastic beds.
Mostly dacitic in composition. Source same as undifferentiated volcanic rocks (Tvnd) but generally
occur more distant from eruptive center. Includes deposits mapped as Upper Quehua Formation.

110 Volcanic rocks, undifferentiated (Miocene and Oligocene).

Chiefly lava flows, but includes extensive pyroclastic deposits and intrusive rocks in some areas
where not mapped separately as units Tig and Ti, and locally may include interbedded nonmarine
sedimenatry rocks. Mostly of andesitic and dacitic composition; sources are poorly defined volcanic
eruptive centers, now deeply eroded. Includes deposits mapped as Abaroa Formation, Bonete
lavas, Camargo volcanic rocks, Carangas volcanic rocks, Cuzco volcanic rocks, Julaca lavas, Mauri
 Formation, Murmuntani Formation, Negrillos volcanic rocks, Pulacayo volcanic rocks, Rondal lavas,
Tagua Formation, and Ubina volcanic rocks.

100 Intrusive rocks (Pliocene to Oligocene).

Chiefly subvolcanic stocks, plugs, and dikes of dacitic composition in vent complex of eroded
volcanic eruptive centers. Not all intrusive rocks mapped; many are included in unit of
undifferentiated volcanic rocks (Tvnd). Important host rock for Bolivian polymetallic vein deposits.

70 Sedimentary rocks (Pleistocene and Pliocene).

Nonmarine sandstone, conglomerate, and shale. May include minor interlayered volcanic rocks.
Includes deposits mapped as Aroma Formation, Charaña Formation, La Paz Formation, Pomata
Formation, Remedios Formation, Topohoco Formation, Taracollo Formation, Ulloma Formation, and
Umala Formation.

60 Ignimbrite (Pleistocene to Miocene).

Welded and nonwelded ash-flow tuffs, chiefly in extensive outflow sheets. Mostly of dacitic
composition. Sources probably large caldera complexes, only a few of which are identified.
Locally, may include basinal lacustrine sediments. Includes deposits mapped as Ignimbrite
Formation and Perez Tuff.

50 Salt deposits (Holocene and Pleistocene).

Playa-lake evaporites. May include interbedded fine-grained lacustrine deposits. Locally may be
seasonally covered with water.

30 Lacustrine deposits (Holocene and Pleistocene).

Chiefly calcareous tufa in ancient lake shorelines and lacustrine mud and silt deposits. Includes
deposits mapped as Minchin Limestone.

20 Stratovolcano deposits (Holocene to Miocene).

Lava flows, flow breccias, lahars, and minor pyroclastic deposits chiefly of andesitic to dacitic
composition. May include domes and shallow intrusive bodies mostly in vent complexes. Includes
deposits mapped as Cerkhe Formation, Estratovolcano Formation, and Huaricunca Formation.

10 Surficial deposits, undifferentiated (Holocene and Pleistocene). Includes unconsolidated alluvial,

eolian, colluvial, and glacial deposits. Locally may include lacustrine and salt deposits that are not
shown separately.

Table 3. Equivalence between GEO1 numeric codes and the geological info.

The new altidis.new_polygon.shp contains 1999 records, grouped in 17 types GEO1 codes,
this means that filling the field DESCIPTION only requires 17 SQL queries like:

UPDATE altidis.new_polygon
SET DESCIPTION = 'Salt deposits (Holocene and Pleistocene). Playa-lake
evaporites. May include interbedded fine-grained lacustrine deposits. Locally may
be seasonally covered with water.'
WHERE GEO1=50

Then, I show two queries for the new file, in two diferent interfaces, Google Earth and
Quantum GIS:
 Figure 3. Query over altidis.new_polygon.kml in Google Earth.

Figure 4. Query over altidis.new_polygon.shp in Quantum GIS 1.7.0-Wroclaw.

 WATERSHED ANALYSIS.

This comprises some phases, first download the SRTM WRS2 data from the Global
Land Cover Facility. The SRTM WRS2 holds elevation data in GeoTiff files, I dowloaded two
files: SRTM_ffB03_p001r073.tif and SRTM_ffB03_p233r073.tif (see the DATA SOURCES
section at the end of this paper for a detailed infomation about this files and its coverage
area).

SRTM_ffB03_p001r073.tif SRTM_ffB03_p233r073.tif

Table 4. Coverage area of the downloaded SRTM WRS2 files.

The geotiff files come in EPSG 32719 (Projection UTM zone 19 South, Datum WGS84),
I did all the watershed analysis in EPSG 32719, only at the end of the work the conversion of
the result files was done, to EPSG 4326 (Projection Geographic, Datum WGS84).

Once the files are downloaded I proceeded to merge them in one file and then write the
result to a DEM format, altiplano.dem. I opened this file with VTBuilder for doing some
preprocessing like fill the holes of missing data. I did this from within the 'Elevation' menu
option: Elevation → Fill in Unknown Areas → Extrapolation via partial derivatives. The result is
shown in the next Figure.

Figure 5. Altiplano_filled.dem.

I did the conversion to GeoTiff for a later processing with GRASS GIS software, this file
is altiplano_filled.TIF. Also I did another conversion, a conversion from DEM data to RASTER
data without elevation infomation, altiplano_filled_shaded_relief_GEO_WGS84.TIF. This
last file is used as the base layer for the representation of all the other layers in Quantum GIS
(See Figure 2).

Figure 6. altiplano_filled@PERMANENT viewed in GRASS GIS.

The watershed analysis begans with the import of the elevation data into GRASS GIS:
altiplano_filled.TIF, the result in GRASS is altiplano_filled@PERMANENT, this's the base
file for the watershed analysis. Next I show the command lines I've employed to do the work.

GRASS command lines:

r.watershed -f elevation=altiplano_filled@PERMANENT accumulation=ACUMULACION

basin=BASINS threshold=400000 convergence=5

r.mapcalc 'LOG_ACUMULACION=log(abs(ACUMULACION@PERMANENT)+1)'

r.mapcalc INF_RIOS = if(LOG_ACUMULACION@PERMANENT>9)

r.thin input=INF_RIOS@PERMANENT output=THIN_RIOS

r.to.vect input=THIN_RIOS@PERMANENT output=rivers_9_utm_wgs84 feature=line

r.to.vect input=BASINS@PERMANENT output=basins_vect_utm_wgs84 feature=area

v.generalize input=rivers_9_utm_wgs84@PERMANENT
output=rivers_9_utm_wgs84_SUAVIZADO method=snakes threshold=1.0 look_ahead=7
slide=0.5 angle_thresh=3 degree_thresh=0 closeness_thresh=0 betweeness_thresh=0
alpha=1.0 beta=1.0 iterations=1

v.generalize input=basins_vect_utm_wgs84@PERMANENT
output=BASINS_VECT_UTM_WGS84_SUAVIZADO method=snakes threshold=1.0
look_ahead=7 reduction=50 slide=0.5 angle_thresh=3 degree_thresh=0
closeness_thresh=0 betweeness_thresh=0 alpha=1.0 beta=1.0 iterations=1
 Figure 7. rivers_9_utm_wgs84_SUAVIZADO@PERMANENT viewed in GRASS GIS.

Once the work has done I proceed to export the two files to SHP files:

SOURCE DESTINATION
rivers_9_utm_wgs84_SUAVIZADO@PERMANENT rivers_9_utm_wgs84_SUAVIZADO.shp →

command line: reprojected to

v.out.ogr -e rivers_9_GEO_WGS84_SUAVIZADO.shp
input=rivers_9_utm_wgs84_SUAVIZADO@PERMANENT
type=line dsn=rivers_9_utm_wgs84_SUAVIZADO
BASINS_VECT_UTM_WGS84_SUAVIZADO@PERMANENT BASINS_VECT_UTM_WGS84_SUAVIZADO.shp →

command line: reprojected to

v.out.ogr -e BASINS_VECT_GEO_WGS84_SUAVIZADO.shp
BASINS_VECT_UTM_WGS84_SUAVIZADO@PERMANENT
type=line
dsn=BASINS_VECT_UTM_WGS84_SUAVIZADO

Table 5. Conversion of the result files to the SHP format.

The last stage of the work is the cleaning of files from the artifacts generated, I did this
with Quantum GIS.
 CUSTOM DATA

I've vectorized some features like lakes, major salt deposits, and the network of rivers
north to Coipasa salt deposit. The mentioned network of rivers isn't complete, I'll complete it
later. I show the result of this work:

Figure 8. A general view of altiplano rivers.kml from Google Earth.

Next I show another more detailed view of this layer:

 Figure 9. A more detailed view of altiplano rivers.kml from Google Earth.

The list of the new vectorized layers is:

Layers for Google Earth Layers for GIS SOFTWARE (Quantum GIS)

altiplano rivers.kml altiplano rivers.shp

Poopo Lake.kml Poopo Lake.shp
Uru Uru Lake.kml Uru Uru Lake.shp
Coipasa salt deposit.kml Coipasa salt deposit.shp
Uyuni salt deposit.kml Uyuni salt deposit.shp
Table 6. List of the new vectorized layers.

I've packed all the KML layers (geologic, watershed and custom) in one KMZ file, see
the DOWNLOAD SECTION below in this paper for download the complete DATASET.
Figure 10. A view of how looks BOLIVIAN ALTIPLANO GIS DATASET.kmz in Google Earth.

Figure 11. A view of the same dataset (SHP files) in Quantum GIS.
 SOME DATA QUALITY RESTRICTIONS

Take in mind the medium resolution of data elevation used as the basis of watershed
analysis, this resolution is 90 meters per pixel.

PROBLEMS WITH QUANTUM GIS

The layer altidis.new_polygon.shp present a problem when a query (info tool) is

done over polygons with the GEO1 field equal to 10 → Surficial deposits, undifferentiated
(Holocene and Pleistocene). Includes unconsolidated alluvial, eolian, colluvial, and glacial
deposits. Locally may include lacustrine and salt deposits that are not shown separately.

When we request information with the info tool over a polygon with GEO1=10
Quantum GIS hangs. I've expended a lot of time working around this problem in Quantum
GIS trying to fix it, but I didn't. Moreover, this problem only appears in Quantum GIS, I've
tested the file against other GIS software (MapWindow GIS, GRASS GIS) without any
problem when I request info of type GEO1=10. Their homologue file in KML format,
altidis.new_polygon.kml didn't present any problem in Google Earth when a query is done
over polygons with GEO1=10. You can view a list of errors detected in
altidis.new_polygon.shp using the optinon menu Vector → Geometry tools → Geometry validity
check.

DOWNLOAD SECTION

Dataset for use with Google Earth:

Download the file BOLIVIAN ALTIPLANO GIS DATASET.kmz (6 MB).

Datasets for use with GIS SOFTWARE (Quantum GIS)

Download the file BOLIVIAN ALTIPLANO GIS DATASET.zip (14 MB).

Download the file BOLIVIAN ALTIPLANO GIS DATASET (with RASTER layer).zip (55 MB).

Figure 12. A overview of 'BOLIVIAN ALTIPLANO GIS DATASET' directory structure. This is
the version intended for use with GIS SOFTWARE like Quantum GIS or others. In this
directory structure you can see the file altiplano_GEO_WGS84.qgs (at the root of the
directory), a project file for use with Quantum GIS 1.7.0 Wroclaw.
 See the SOFTWARE USED section at the end of this paper with links to the download
pages of that software.

DATA SOURCES

• From USGS Geoscience Data Catalog:

Digital Geologic Map of the Altiplano and Cordillera Occidental, Bolivia

The original geodata belongs to:

Originator: Sherman P. Marsh, Donald H. Richter, Steve Ludington, Eduardo

Soria-Escalante, Angel Escobar-Diaz.
Publication_Date: 1995.
Title:
Digital Geologic Map of the Altiplano and Cordillera Occidental, Bolivia
Geospatial_Data_Presentation_Form: map.
Series_Information:
Series_Name: U.S. Geological Survey Open-File Report.
Issue_Identification: 95-0494
Online_Linkage: <http://pubs.usgs.gov/of/1995/ofr-95-0494/>

• SRTM WRS2 data from GLCF (Global Land Cover Facility).

Granules used:

granule_id: 173385
online link: http://glcfapp.glcf.umd.edu:8080/esdi/ftp?id=173385
ID: 173-385
PATH-ROW: 001-073
ATTR: 3 arcsec, Filled Finished-B.
ZONE PREVIEW:

granule_id: 181731
online link: http://glcfapp.glcf.umd.edu:8080/esdi/ftp?id=181731
ID: 181-731
PATH-ROW: 233-073
ATTR: 3 arcsec, Filled Finished-B.
ZONE PREVIEW:
SOFTWARE USED

Virtual Terrain Project home page (VTBuilder, Enviro, Cmanager).

Virtual Terrain Project download page.

Quantum GIS 1.7.0 Wroclaw home page.

Quantum GIS 1.7.0 Wroclaw download page.

GRASS GIS v 6.4.1 home page.

GRASS GIS v 6.4.1 download page.

REFERENCES

(1) A newGRASS GIS toolkit for Hortonian analysis of drainage networks.

(2) Watershed analysis.
(3) R.stream.* GRASS.
(4) An approach to the hydraulic culture on the Bolivian Altiplano.

EMAIL CONTACT INFORMATION

Email: cgp2258@gmail.com

WEB PAGE AUTHOR'S

European West Trail.