LSP: expand points to areas from WDPA for MPAs

Hi Courtney,

Given area and location of Ocean Tipping Points, you can certainly create polygons of circles per location.

First, you'll want to calculate the radius (r) given the area (a): a = pi*r^2; r = sqrt(a/pi). Then you have a table, for example with fields: id, name, lon, lat, area_km2, radius_km. In ArcGIS, you would generate points from the table with Make XY Event Layer, and Buffer these points with the radius_km field. In R, you would read.csv, sp::coordinates to create points and apply them to rgeos::gBuffer.

For a more nuanced approach restricting the area to only habitats (ie clipping land), attached and below is the paper in which they converted WDPA points to habitat specific (ie terrestrial vs marine) areas.

Visconti, P., M. Di MARCO, J. G. áLvarez-Romero, S. R. Januchowski-Hartley, R. L. Pressey, R. Weeks, and C. Rondinini. 2013 Effects of Errors and Gaps in Spatial Data Sets on Assessment of Conservation Progress: Errors and Gaps in Spatial Data Sets. Conservation Biology

This paper borrows from:

Jenkins, Clinton N., and Lucas Joppa. 2009**. Expansion of the Global Terrestrial Protected Area System. Biological Conservation

with this blurb in particular from Visconti:

To explore the implications of using buffered protected- area centroids when the actual shape of PAs is unknown, we followed a similar approach to that of Jenkins and Joppa (2009). We replaced protected-area polygons with their buffered centroids (circles equivalent in size to the actual polygon and centered on the polygon centroid) to assess the protection of habitat for Neotropical ter- restrial mammals (n = 1558), derived from Rondinini et al.’s (2011) habitat suitability models. In contrast to the ecoregional analysis of Jenkins and Joppa (2009), our biodiversity data were at a much finer spatial resolution (300 m).

I'm friends with Jenkins and Joppa from Duke days, so happy to facilitate a convo and/or request code. Worth looking up forward citations from Visconti et al, and Jenkins & Joppa.

PS Dropped these references into zotero OHI / LSP library, and creating ohicore issue #77.

---------- Forwarded message ---------- From: Ben Best bbest@nceas.ucsb.edu Date: Tue, Aug 27, 2013 at 7:08 PM Subject: Re: WDPA data Qs To: Ben Halpern halpern@nceas.ucsb.edu Cc: Katie Longo longo@nceas.ucsb.edu

Ok, going ahead with just Designated makes sense to me. I see some Ramsar sites that are not otherwise covered by national parks (but most are). Here's my tentative plan to finish tonight:

Filter WDPA polygons by "STATUS"='Designated'
When converting polygons to raster based on "STATUS_YR", I'll prioritize "DESIG_TYPE"='National', and the earliest year.

I'll hold off on buffering the points. This paper attached goes into excellent detail on the merits and downfalls of this sort of analysis using coral reef cover as an example. I'd like to consider adding their recommendations to our TODO for next year:

We recommend end users (1) clip the portion of buffered centroids of PAs known to be only terrestrial or marine to their respective realm; (2) clip the portion of buffered centroids of PAs known only from one country to the country borders;...

Thanks, BB

Here are counts on "DESIG_TYPE":

International      National 
        28004        145378

And here are counts on "DESIG_ENG" of the "DESIG_TYPE"='International':

 Freq                                                             DESIG_ENG
    26                                                  ASEAN Heritage Park
   159                                   Baltic Sea Protected Area (HELCOM)
     1                                       Marine Protected Area (CCAMLR)
     6                                        Marine Protected Area (OSPAR)
  1214                     Ramsar Site, Wetland of International Importance
 20892                    Site of Community Importance (Habitats Directive)
  5244                            Special Protection Area (Birds Directive)
    20 Specially Protected Area of Marine Importance (Barcelona Convention)
   219                                         UNESCO-MAB Biosphere Reserve
   223                                                  World Heritage Site

On Tue, Aug 27, 2013 at 5:51 PM, Ben Halpern halpern@nceas.ucsb.edu wrote: Hi BB, Don't add the extra ones from Marla - those are now included in WDPA so are duplicates. I agree to remove the international ones but think we should only use designated (inscribed usually means paper park). Same with adopted. I lean towards only doing polygons. We thought through this issue a lot last year so feel good about sticking with it. Ben

Benjamin S. Halpern Director, CMAP Research Biologist NCEAS, UC Santa Barbara

On Aug 27, 2013, at 5:18 PM, Ben Best bbest@nceas.ucsb.edu wrote:

Hi Ben and Katie,

The WDPA has a STATUS field, most of which are "Designated" but a few other fields which are less clear. Here are the counts for the 5 categories:

     Adopted   Designated    Inscribed Not Reported     Proposed 
          20       170769          223          297         2073

The official metadata document attached only mentions 'Designated' (want) and Proposed (don't want). We previously used only "Designated".

In the other paper attached, they use all but Proposed, which I advocate. They also removed the largely duplicative "international" reserves and buffer the WDPA points, both of which I could do today if you think reasonable. Here's their methods:

We used the December 2012 version of the WDPA for analysis (IUCN & UNEP, 2012). The WDPA is provided as two separate GIS shapefiles: ‘WDPA polygons’ for protected areas where the boundary and shape of the protected area is known, and ‘WDPA points’ for protected areas where only the point location is known. Where sites only existed in the WDPA as a point location, we used the ‘buffer’ tool in ArcGIS to create a circular polygon of the same size as the given area of the protected area (as recorded in the WDPA), with the point location as its centroid. We then used the ‘Merge’ tool to add the buffered points to the existing WDPA polygon shapefile. We included protected areas with a designation status of ‘adopted’, ‘designated’, ‘inscribed’and ‘not reported’, and excluded ‘proposed’ protected areas. All reserves with international designations(World Heritage, Ramsar and Man and Biosphere) were removed leaving only nationally designated reserves, as most international designations either duplicate national reserves or may not meet the requirements for full protected area status (selection of nationally designated areas has also been applied in previous analyses of protected area coverage: see Jenkins & Joppa, 2009, and Schmitt et al., 2009, among others). The final version of the WDPA for analysis contained 168,054 nationally designated protected areas, of which 12 per cent were buffered points.

Here's Darren's dataedit/ingest/GL-WDPA-MPA/README.txt, which also refers to 8,965 designated protected areas that look like they were downloaded by Marla based on prefix mranelletti-search*. Should I add those too or do a spatial check for duplication with latest WDPA?

We extracted vector data for 98,358 designated protected areas from the WDPA (2010) database (using STATUS="Designated"), and projected into Mollweide. Each record for a designated protected area includes the year in which the protected area was designated (STATUS_YR), ranging from 1872-2008 and NoData (0). We also extracted an extra 8,965 designated protected areas by downloading vector data from the WDPA website for select countries who were not included in the WDPA (2010) database (for various reasons), and then we excluded areas without explicit spatial area data (IS_POINT=1):

Dominican Republic (no designated areas)
Gambia (2 designated areas)
Guinea (106 designated areas)
Singapore (7 designated areas)
United Kingdom (8,849 designated areas)
Uruguay (1 designated area)

We rasterized all the WDPA designated areas using 1km resolution and the year designated as the cell value, resolving conflicts using the oldest year as a priority field and excluding NoData -- e.g., $1-(year/2008)$ where $year>0$.

References:

IUCN and UNEP. 2010. The World Database on Protected Areas (WDPA). UNEP-WCMC. Cambridge, UK. www.protectedplanet.net
United Nations Environment Programme World Conservation Monitoring Centre (WCMC). Protected Planet: The Protected Areas Portal. Data Standards for the World Database on Protected Areas. BETA VERSION 1. January 2010.

Technical notes: These data have a flag for MARINE with values of true, false, and Not Reported. A full twenty percent of the protected areas have Not Reported for their MARINE flag. When we extract the data using MARINE=true, we find about 4,800 protected areas, many of which are most than 10km offshore (purposefully due to management boundaries near shore).

Thanks, BB <WDPA Data Standards 2012.pdf>

OHI-Science / ohicore

LSP: expand points to areas from WDPA for MPAs #71