New Hampshire/Conservation Clean Up

Between 2009 and 2011, a large amount of data was imported from GRANIT. One of those imports was the conservation layer. It was originally imported using landuse=wood. Several years later, it was changed to natural=wood. There are several reasons why natural=wood is not a good tag for this data:

• This is cadestral data, and parcel lines rarely line up with land cover. This leads to situations where swamps, lakes, rivers, farms, and parks are within natural=wood areas.
• This use of natural=wood is not the documented definition of natural=wood.
• Mappers who are unfamiliar with this use of natural=wood treat it as if it were land cover and make multipolygon relations that do not follow the conservation area boundaries.
• People complain about uses of natural=wood that follow the documented definition. Since most uses of natural=wood in the state currently mean "Conservation Area", people complain about uses of natural=wood that mean "wooded land cover".

This project will use the conservation dataset available at the GRANIT website in order to perform an informed clean-up of the import. The plan is to have this page be a repository of recommendations for the clean-up project that mappers can start working on at any time.

A stated non-goal of the project: conservation areas that were added in the last decade will not be imported. Mappers are welcome to use the data to add new conservation areas separate from this project, though of course not in bulk and with community consultation to ensure appropriate tagging.

Available Information

This section lists some of the information that's available in the GRANIT data to use. Each tag in the original data will provide some information as to how the way should be tagged in OSM. Not all of them will be are important for this project, but they should be documented nonetheless.

PPTYPE

This tag says what kind of conservation area an object is. There are options for agricultural restrictions, historic preservation, scenic easements, fee ownership, and several others. It appears that the original import only brought in the FO (fee ownership), CE (conservation easement) and SE (scenic easement) categories, though some areas were reclassified in the last decade. Areas in the scenic easement category should be considered for removal from OSM.

PPTERMTYPE

This tag says whether the parcel is protected indefinitely or for a limited time. This isn't of interest for this project, but other mappers may want to add protected=perpetuity to objects with the U value.

PPAGENCY/PPAGENTYPE

This tag specifies the protecting agency of the land. This isn't of interest for this project, but other mappers may want to add operator=* based on these values. The original import seems to have only included the conservation easement areas that were protected by a government agency (i.e., did not include those with PPAGENTTYPE=5).

LEVEL

This tag designates how strongly the land is conserved. A value of 1 indicates that the land is expected to be in a natural or quasi-natural state. A value of 4 indicates that the land can be used for anything. Unfortunately, the data quality for this tag is rather poor. Objects with values of 1, 2, or 3 could be tagged with leisure=nature_reserve and/or boundary=protected_area. Objects with a value of 4 either were not imported or have already been correctly tagged, so they will not be a concern for this clean-up project.

MSTATUS

This tag indicates the management status of the protected area. This can be translated into boundary=protected_area + protect_class=* tagging. Here is a guide:

• 1 => protect_class=1b in a wilderness area, protect_class=5 otherwise
• 2 => protect_class=5
• 3 => protect_class=6
• 4 => No protect_class

OWNERTYPE

This tag identifies whether a conservation is on state, federal, local, or private land. ownership=* tagging could use these values:

• 1 => ownership=municipal
• 2 => ownership=national
• 3 => ownership=state
• 5 => ownership=private
• 6 => ownership=county

Recommendations

• Because the original import excluded objects in the GRANIT data that do not work well with the definition of leisure=nature_reserve, it is recommended that all imported ways currently tagged as natural=wood be changed to leisure=nature_reserve. While the leisure=nature_reserve tag could potentially be incorrect for some objects with MSTATUS=3 (significant extractive use is allowed so long as >50% of the land cover is natural), that basically only applies to the ski resorts, which have already been corrected.
• Ways where the protecting agency (PPAGENTYPE) is public but the land ownership (OWNERTYPE) is private should be considered for removal based on street-level imagery and information from the protector's or owner's websites. Conservation areas that are not publicized can be removed.
• Ways with the scenic easement PPTYPE (SE) should be removed.
• Geometries should be updated and sections that were previously split should be merged into single ways or relations. The new geometries should be vetted, of course. They are known to be inaccurate around roadways and streams especially.

The Numbers

As of Sunday, 14 March 2021, there are 1227 ways and relations with source=GRANIT and natural=wood. This matches to 1223 features in the modern data. 1062 (86.8%) of these ways have PPTYPE=FO. There are 16 features that are tagged with ACCESS=3 (private). There are 74 features with a public protecting agency (PPAGENTYPE != 5) and a private landowner (OWNERTYPE == 5). There are 25 scenic easements. Lastly, there are only 14 features where the geometries match 100%. In total, there are 177 features that need a closer look to check whether the nature reserve tagging is correct.

Action Plan

• Saturday, 13 March 2021: Share original draft of this document with OSMUS community Slack workspace for initial feedback and collaboration
• Sunday, 14 March 2021: Calculate exactly how many ways would be effected to help decide whether it would be worth starting a MapRoulette challenge or doing it all by hand given that the vast majority would be a simple tag change, and edit this document to reflect that
• Afterwards: Start working on the largest, most complicated areas like Mount Monadnock, Mount Cardigan, and the Andorra Forest. Then start working on smaller, simpler areas, which will be much faster