1.0

The dataset is captured as part of CapeNature’s fire management policy to provide an accurate record of the fire history of those areas managed by CapeNature, in order to facilitate ongoing analysis of veld age, fire frequency and fire return intervals as indicators of the state of the veld managed by CapeNature.

All data are supplied without distribution rights. You may not redistribute, copy or sell in any form (value added or not) these data to any other parties/bodies outside your organisation/department. All data sets are supplied with metadata files which, inter alias, include information on the custodian of the specific data set. Any reports and/or publications resulting from your project must acknowledge the Western Cape Nature Conservation Board (WCNCB), Scientific Services and, if different, the custodian of the data.

All data are supplied with no expressed or implied warranty as to their suitability for purpose, planimetric accuracy or completeness. The WCNCB cannot be held responsible for any errors, which may occur in provided data sets..

HISTORY: Digitising of historical fire records was initiated by Chris Burgers in 1998, with the assistance of Riki de Villiers. It was substantially added to during 2006-7 by the CAPE GEF Fire project by Granville van Ross and Belinda Munsami. All this work was vetted by and large numbers of further records added by Arne Purves, Helen de Klerk, Cher-Lynn Petersen and Therese Forsyth during 2006-9. Significant contributions were also made by Anne-Lise Schutte-Vlok, Paul Buchholz, Tony Marshall and Patrick Meyer. Our thanks to Armin Seydack who made many fire records available for various reserves in the Southern Cape, and Greg Forsyth making records available for fires on state forests land, that were not available in CapeNature’s records.

Since 2000, most Reserve Managers digitize fires themselves using on-screen digitizing over aerial photos and satellite images and GPS (either by walking or flying the fire line) and enter the attributes of the fire report directly into CapeNature’s fire database. All records are centralized and collated by the GIS Fire Technician at Scientific Services.

During a follow-up project, extensive data cleaning was done by Therese Forsyth from July 2012 to 2018 to topologically clean the fire scar shapefiles. This entailed overlaying fire scars per fire season and cleaning out duplicates, slivers and overlaps that lead to too short fire intervals. For fires since 2005, the fires were also verified against the annual SPOT5 satellite mosaic images that were made available on an annual basis and since 2016 using satellite imagery made available online. Where possible fire scars were redigitized based on these satellite images.

For the past 5 to 6 years, many large fires were digitized from downloaded georeferenced thumbnail satellite images of Landsat 8 or Sentinel-2. The thumbnails are available from the USGS website.

The All_Fires_23_24_gw.shp dataset was created by merging all the individual fire shapefiles into one fire shapefile using “Geoprocessing | Merge”. This merged shapefile already contained a column called [FIRE_CODE] as all the individual shapefiles were standardize to have this column. Some basic checks were performed to ensure all individual shapefiles were merged, such as comparing it with the dataset generated in the previous year (using table “Joins”) and checking fires reported in past fire season against the fire register.

The columns [Month] (short integer) and [Year] (long integer) were added and the information for these columns were extracted from the fire code using the “Field Calculator, Python” and the script-code [Month] = !FIRE_CODE![5:7] and [Year] = !FIRE_CODE![8:12]. The columns [Res_Code] (text, 5) and [Land_Unit] (text, 50) were added. The [Res_Code] field can be populated using the “Field Calculator, Python” and the script-code [Res_Code] = !FIRE_CODE![0:4]. The [Land_Unit] information was derived from the ”CNC_Landscape_units.shp” data layer by (i) running an “Analysis Tools | Overlay | Identify”, (ii) then a “Geoprocessing | Dissolve” on [Res_code] with field statistics “Max” of [LUnit] field, and (iii) then “Join” this table to the merged shapefile so that the [MAX_LUNIT] values can be copied across to the [Land_Unit] field.

Further attribute information was extracted to MS Excel spreadsheets from the Central Fire Database. The tables extracted were tblFire.xlsx, tblIgnition.xlsx, and tblIgnitionCause.xlsx. These table were then converted to CSV (comma delimited) files and loaded into ArcGIS. First (i) joined “tblFire.csv” to “All_fires_23_24.shp” using fields [FireReference] to [FIRE_CODE] and exported it to “All_fires_temp1.shp”, (ii) then joined “tblIgnition.csv” to “All_fires_temp1.shp” using fields [FireID] to [FireID] and exported it to “All_fires_temp2.shp”, and (iii) then joined “tblIgnitionCause.csv” to “All_fires_temp2.shp” using fields [IgnitionCauseID] to [IgnitionCauseID] and exported it to “All_fires_23_24_all_attrib.shp”. Deleted the unnecessary fields (Firerefere, Firespotte, cnareabur, mcaareabur, otherareab, totalareab, reportcomp, approvedbu, qualityche, featurecou, firespot_1, notes, nearmiss, reservecen, postfiremo, ignitionid, fireid, dds, dde, comparmen, veldtype, veldage, altitude, landowner, tblignitio, causerelia, otherdescr). The layer was then spatially sorted ascending by year using “Data Management tool | General | Sort”. Renamed the file back to All_Fires_23_24_gw.shp.

The area in hectare was calculated in ArcGIS.

SANS 1878

FGDC-STD-001-1998