We went to the State of the Map US 2013 conference this weekend in San Francisco.
This is the annual conference for the Open Street Map project (OSM)
We (myself and Laci) attended this conference because we view Open Street Map as an important project with a lot of potential for how we can improve public understanding of water in California.
To get an understanding of how active Open Street Map is, Mapbox created this (gorgeous) annual data report Thousands of people around the global actively edit this map of the world, which is primarily focused on streets, and provides a base map experience for the likes of FourSquare and many, many community organizations.
We were not sure about what our opportunities might be with respect to the New California Water Atlas and Open Street Map. We talked to a bunch of people about our project, and met a number of people interested in mapping and hydrology.
If OSM had an initiative to improve water-related map data overall, then it would be possible to deliver constantly improving map tiles far and wide.
It would be possible to relate river, lakes and streams ("surface water") to various water-related features (such as stream gauges, aqueducts, water towers, dams, bridges.)
When we spoke with one of our project mentors, Seth Fitzsimmons from Stamen, he suggested that we could figure out how to build up OSM & the National Hydrography dataset, and have a dialogue about this at the SOTMUS conference. So, we convened a group on Sunday afternoon to talk, and subsequently learned a lot about the various obstacles and opportunities.
Sidenote: Vector tiles
There was one Birds of a Feather session, before ours, about vector tiles with folks from all web-based mapping platforms in attendance. Vector tiles will be highly efficient ways to get maps for higher zoom levels at lower cost. They are fast and will be very easy to color and restyle.
Our big interest is how we can have vector tiles for water. If we could flip on and off certain kinds of features, then we could have a map with up to date water as well as contextual information.
The importance of water to the landscape.
I was personally baffled to learn that not everyone believes water to be important, though many people do.
Rather than beat one's head over debates about what Open Street Map is and isn't, or what a map is and isn't and who gets to define that, let's skip ahead and deal with some realities and possible solutions. We do maintain the assumption that accurate maps are better for everyone in this context.
- Rivers move In 5-15 years a river can, and often does, shift position.
- Restoration efforts happen, which add new water features, bit by bit.
- With climate change, the water landscape may shift even faster, as we have a simultaneous need to communicate about the water.
- Water presents an obstacle to pedestrians. Aaron Ogle's presentation of walkshed.js uses water as an obstacle to predicting a walkshed on top of OSM data. We tend to not walk through water, we tend to go around it.
- Water is a point of interest.
- Transportation. Water can be a 'road' itself - as rivers, or canals.
- Not all rivers are mapped. Gregor, an anthropologist who works on the ClearWater project spoke about the need for better river data in the Amazon, where rivers are how people get around and are essential to a place.
- Water is a place of significance in many communities, or could be.
- In remote communities, there is less infrastructure for mapping water, and so the resources provided through OSM can be of great value.
- Some remote places, such as certain National Parks, have missing water data, and might benefit from community mapping.
Getting accurate water data into and out of Open Street Map
- Rivers in areas covered by trees can't be seen from aerial maps. Example, the Pacific Northwest.
- It is possible to get accurate surface water information into OSM by tracing rivers on aerial maps, or importing GPS information from walking or kayaking in a river.
- OSM encourages getting outside to do the mapping.
- We will assume it is totally OK to use the ID editor.
- We will be able to export changes made to OSM to water features if and when we ever need to do a comparison between government surveyed and citizen mapped waterways. Eric Fischer from Mapbox offered to help us do an "extract" when and if we ever get to that point.
- The Forest service sometimes needs to use their expert knowledge to prune the data, removing streams and springs that no longer exist.
- Imports & OSM = Watch out! Doing an import from an existing high quality dataset has happened. For example, in the United States one of the main datasets for stream location is the National Hydrography Dataset, which is sometimes considered 'too rich' for Open Street Map, it was already imported in places. There is also the TIGER dataset which has some water information as well. In the Open Street Map community, bulk imports are a touchy subject.
- Specialists actually do a really good job of surveying water, in places that have water surveyors.
- Inaccurate data can kill. There were a bunch of great presentations about the US National Parks and OSM. One thing that happens is that inaccurate information (of roads) can cause someone using GPS to get really really lost, and there was a situation where a mom with a baby got trapped and the baby died and so the NPS takes accuracy seriously. Perhaps there are similar needs and expectations with respect to water, my best guess would be a hiker looking for water on a map, but I'm not a good hiker and base most of my knowledge of this from Man vs. Wild, so do not quote me on this. Someone please advise us so that us amateur geographer/cartographers can know what matters.
- We are still not sure how to get information out of OSM.
Possible areas for improvement
- Water labels in OSM might be lacking/not descriptive enough.
- We are invited to give feedback on the new ID Editor for improving water feature information.
- Apparently, it is totally fine to trace a river that you see on an aerial map that does not have geographic information. It is also likely that there is more subtlety to how this actually works.
- Watersheds? To intelligently describe water, watersheds matter. A watershed is actually a description of the way that water falls from an elevation, and is gravity based and not politically based.
According to community legend, the OSM community at some point decided that the project would never include watersheds, even though it has administrative boundaries. We would like to find this statement in print.
- Accept the realities and limitations of OSM:
We could, if absolutely necessary and with enough support, fork our own Open Water Map, or even an Open Eco Map with http://geogit.org. We are not sure that anyone really has the energy or capacity to do this. There might be other kinds of data that make sense as allies, and it would requires explaining what our goals were.
It's possible that water really does have its own special needs. Water does not always have a direction. Water goes into the Earth and out of sight and that matters. Some water is tidal. Do we really need water's directionality for OSM if we just want to be able to create a map that describes water features?
We might need to come up with a list and communicate which kinds of things could go well in OSM with respect to water. This could include a limited list such as streams, rivers, lakes, dams, aqueducts.
For features and tags, deciding what kinds of features would be useful or even possible. Stream Gauges? Sewer covers?
What kinds of tags would be helpful or are even possible?
Community Mapping opportunities
A number of ideas were presented for how water information could be improved. Since Open Street Map is volunteer-driven, building capacity to improve water will require a dedicated group of people to work on this and good community organizing.
Email list? We have an email list of people in our group, and also would like to join whatever mailing list is appropriate. If someone can direct us to a water related mailing list for Open Street Map that would be great, we will continue this discussion there where it will be most productive.
Work on a watershed by watershed basis. Since OSM is a community focused project, it would be easier, better and possibly more accurate to organize watershed improvement projects. This would be a better way to deal with moderately sized imports and ensure that a community is available to verify the map and ensure and maintain accuracy.
Create opportunities to use OSM for creek restoration projects Jeanny has a project in San Leandro, CA.
Organize around mapping urban watersheds In cities, water changes. (How, I'm not sure, possibly because of development?) There is also enough population density to have people work on water for their place of residence.
Kayaking trips for real!
I will update this post with new information as necessary. Please comment below.
These are the people who came to our discussion and contributed to the ideas presented here.
** Please comment if I missed anything or got things wrong **
- Daniel Spring
- Laci Videmsky
- Seth Fitzsimmons
- Chacha Sikes
- Gregor Maclennan
- Richard Burcher
- Rob Brackett
- Paul Norman (OSM user, did NDH import, pnorman)
- Tamara Colby