[Talk-ht] River system analysis tool for Haiti

Pierre Beland infosbelas-gps at yahoo.fr
Mon Oct 25 18:48:29 BST 2010


The cholera epidemy and water contamination shows the importance of well identifying the Drainage Bassins in Haiti. These like a tree leaf, have complex ramifications not easy to analyze. I am not a GIS specialist and dont know if specialized tools exist to analyse such ramifications.

Here are some observations following the validation work I did manually last spring on Haiti Rivers, streams and canals.

The affluents of a river are complex ramifications.  See, for example, http://www.openstreetmap.org/?lat=18.4057&lon=-72.0741&zoom=14&layers=M . In such a ramification, its long and tedious to check each ramification, to validate if directions are appropriate.

To this regard, the Report http://osm3.crans.org/haiti/suivi-affluents.php shows only small part of Drainage bassins.

To analyse and build good cartography of Drainage bassins, we have to consider all the waterways and all ways and nodes connected to. I have identified waterways with the following tags : river, stream, canal and drain.

I have often seen ways with the wrong direction (going upward).  Yan report on rivers also shows many small waterway segments not connected to a Drainage Bassin.

We have to consider the possibility that ways are connected to a River ramification with wrong tags or are wrongly connected to rivers.

Ideally, we would have to analyse more deeply the river ramifications and identify different incoherencies, check if all river names are present.

If useful to identify steep slopes, it could also be possible to add elevation to each node.

To do efficiently the validation work, it would be necessary to extract all the nodes and ways connected to river ways and relations.

And badly, that, I dont know how to do with Openstreetmap database. If it was possible to easily export all the information related to rivers in a MySQL database, I think that we could do a lot of validations from that database.

In Yan List of Rivers (http://www.tappenbeck.net/osm/data4wiki/haiti/haiti-river_wiki.html ), there are still many unnamed segments. Many of them might be unconnected to other rivers. This list have to be manually updated and might have not been updated recently.

Opencyclemap contour lines could be usefull to identify river directions. But rivers are not showed on that map.  I dont know if there are ways to add simultaneously contour layers and river layers to Openstreetmap.

Exemples of problems
Canal Desdunes (waterway=canal) not connected page http://www.openstreetmap.org/?lat=19.2867&lon=-72.5855&zoom=14&layers=M 

Many River ramifications not connected http://www.openstreetmap.org/?lat=18.4397&lon=-71.9705&zoom=14&layers=M

A stream connected to a road http://www.openstreetmap.org/?lat=18.4579&lon=-71.8602&zoom=14&layers=M 

Rivers ending nowhere (or in canals ?) in suburb of Port-au-Prince http://www.openstreetmap.org/?lat=18.5218&lon=-72.125&zoom=13&layers=M 

When looking at zoom level 13,14, you will see well the ramifications and see a lot of problems with orphan rivers.

Pierre Beland 

De : Jean-Guilhem Cailton 
Date/heure : 2010-10-23  19:34:56 
A : HOT; Talk-ht 
Cc : OSM-talk; Jocelyn Jaubert 
Sujet : Re: [Talk-ht] River system analysis tool for Haiti 
Jocelyn has setup a page showing the tributaries network for Haiti, 
updated from hourly updates if all works well:
For waterways to appear in it, they have to be represented by relations, 
and tagged according to the scheme described in:
i.e. with tributaries appearing as members of the destination relation, 
with role=tributary, and being themselves relations if they have their 
own tributaries.
(I have just learned how to add a relation as member of another relation 
in JOSM:
- double click the tributary relation in the relations panel, it should 
appear in pink in the main editing area;
- single click the destination relation;
- click on the "Modify relation" button).
It is also considered to maybe adapt the page
(originally authored by Sly, and currently using ways only, not relations)
to build the network from topology information and display it in a table.
Many thanks to Jocelyn (and Frodrigo for discussion).
News from biosurveillance do not sound good, with
"Confirmation of Cholera at Multiple Sites in Port au Prince"
An extract from yesterday IFRC Information bulletin:
"A team of epidemiologists from the Department of Public Health and 
Population (MSPP) was mobilized to
Artibonite to analyze samples taken from several patients to identify 
the causative agent of the acute
diarrhoea outbreak. As a result of these analyses, the Ministry 
confirmed that Vibrio cholerae is causing the
outbreak and that the source of the contamination is the river."
Talk-ht mailing list
Talk-ht at openstreetmap.org
-------------- section suivante --------------
Une pi?ce jointe HTML a ?t? nettoy?e...
URL: <http://lists.openstreetmap.org/pipermail/talk-ht/attachments/20101025/9e0ed3c0/attachment-0001.html>

More information about the Talk-ht mailing list