OR/15/021 Technical information

From MediaWiki
Jump to navigation Jump to search
Wildman, G, Adlam, K A M and Cartwright, C E. 2015. User Guide for the GeoSure Insurance Product (version 7). British Geological Survey Internal Report, OR/15/021.

Hazard: Exposure or vulnerability to injury or loss.

Risk: The impact of the hazard on people, property or capital.

For example, a flight of stairs could be perceived as a hazard, but the likelihood of falling down them would be the risk.

A high hazard does not necessarily translate to a high risk. For example, if a particular location has a relatively high ground instability hazard, but the properties that are built there have taken this into account, and are designed to withstand the hazard, they will not have a comparable level of risk. This is because the likelihood of the hazard causing any loss has been reduced due to the design of the property.

GeoSure does not identify the cost of a hazard being realised, and therefore does not consider risk. GeoSure only examines the conditions that leave an area exposed to a hazard. It is therefore a hazard dataset.

Scale

The GeoSure Insurance Product (GIP) dataset is produced for use at 1:50 000 scale providing 50 m ground resolution.

Field descriptions

Table 1 GeoSure Hazard ‘Visualisation’ layers
Field name Field description
TILE Ordnance Survey grid square
ORIG_FID Unique Feature ID
VERSION Dataset name and version number
Table 2 GeoSure Derived Postcode Database
Field name Field description
POSTCODE Postcode
COLLAPS_OF Collapsible occurrence factor
COMP_OF Compressible occurrence factor
DISSOL_OF Soluble rocks occurrence factor
RSAND_OF Running sand occurrence factor
SLOPE_OF Landslide occurrence factor
SSWELL_OF Shrink swell occurrence factor
TOTAL_OF Sum of the six hazard occurrence factors for each postcode
CLASS5 The Total_of column is classified into 5 classes using natural breaks (Jenk’s Classification), values of 1–5
CLASS10 The Total_of column is classified into 10 classes using natural breaks (Jenk’s Classification), values of 1-10
VERSION Dataset name and version number

Creation of the dataset

GeoSure Hazard ‘Visualisation’ layers

The six HV layers are directly derived from the master GeoSure hazard layers as follows:

  • Each of the six master GeoSure hazard layers is already classed into 5 classes A–E (see Table 3). Classes D and E are selected from each of the master layers to create six new Hazard Visualisation (HV) layers.
  • The polygons in each of the new HV layers are dissolved to remove boundaries between classes D and E to improve performance and reduce complexity.

GeoSure Derived Postcode Database

In detail, the methodology behind producing the DPD is as follows: The postcode centroids layer (derived from Ordnance Survey Code-Point® Open data) is buffered using a value of 300 metres. This value takes into account the variability in size of postcodes (250 metres) and the 50 metres cartographic precision of the Digital Geological Map of Great Britain at the 1:50 000 scale (DiGMapGB-50) data that is used as the geological basis for the GeoSure hazard layers.

The methodology involves balancing (normalising) the 6 GeoSure natural ground stability hazards against each other in relation to their spatial coverage. The GeoSure maps themselves have a fivefold coding (A to E), and the balancing exercise involves comparing each level across the six hazards e.g. comparing a level C shrink-swell clay area with a level C running sand area. The comparison is done by estimating how frequently a ground movement event may occur that could have the potential to damage a property (in particular, low-rise buildings). Each level of each of the hazards is given an ‘occurrence factor’ (as shown in Table 3), which can then be added together to derive a Total Occurrence Factor (TOF) at a particular location (e.g. within a given postcode).

Table 3 Normalisation table for GeoSure hazards applied to low rise structures for the insurance industry.
Landslide Compressible Running sand Shrink-swell Soluble rocks Collapsible
Class Nominal score Occurrence rating Unified rating Occurrence rating Unified rating Occurrence rating Unified rating Occurrence rating Unified rating Occurrence rating Unified rating Occurrence rating Unified rating
A

100

0

0

0

0

0

0

0

0

0

0

0

0

B

100

0.001

0.1

0.001

0.1

0.001

0.1

0.001

1

0.001

0.1

0.001

0.1

C

100

0.002

0.2

0.002

0.2

0.001

0.1

0.02

2

0.002

0.2

0.001

0.1

D

100

0.01

1

0.01

1

0.002

0.2

0.1

10

0.005

0.5

0.002

0.2

E

100

0.02

2

0.1

10

0.005

0.5

0.2

20

0.01

1

0.005

0.5

Key to definitions.

Class The coding system for the 6 GeoSure natural ground movement hazards shown in the column headers
Nominal score Numeric representation of the Class Codes in order to allow combination with the occurrence rating
Occurrence Rating The estimated frequency of the hazard potential being realised in terms of damage to buildings with shallow foundations (less than 2 metres deep). Table 4 below shows the scale used for the occurrence rating.
Unified Rating The result of the nominal score times the occurrence rating.
Table 4 Calibration of the hazard Occurrence Rating.
Estimated Frequency of Occurrence Numerical Representation of Estimated Frequency of Occurrence Occurrence Rating
once in 1 years 1/1 1
once in 5 years 1/5 0.2
once in 10 years 1/10 0.1
once in 20 years 1/20 0.05
once in 50 years 1/50 0.02
once in 100 years 1/100 0.01
once in 200 years 1/200 0.005
once in 500 years 1/500 0.002
once in 1000 years 1/1000 0.001

Before use in the DPD, these unified rating values in the table are then multiplied by 10 to allow the use of integer values. For example a class 'D' shrink swell pixel will have a value of 10 x 10 = 100. This is to enable geo-processing operations to compute weighted averages for each postcode area.


Therefore the ranges of the data in the DPD are as follows:

Collapsible Deposits Occurrence Factor: 0–5 Compressible
Ground Occurrence Factor: 0–100
Soluble Rocks Occurrence Factor: 0–10
Running Sand Occurrence Factor: 0–5
Shrink-Swell Occurrence Factor: 0–200
Landslides Occurrence Factor: 0–20

Although there is a theoretical maximum TOF of 5+100+10+5+200+20 = 340, the maximum score achieved under existing geological conditions is 213.

  • Using a GIS, the Total Occurrence Factor (TOF) values for each hazard are converted to 6 grids representing each of the 6 GeoSure layers.
  • In order to assign a score to each postcode unit, statistics are calculated using the buffered postcode layer in conjunction with each of the six TOF hazard feature layers. This process considers each hazard in turn and calculates the average values proportional to their spatial coverage, that occur within each of the 300m circular buffered postcode areas. These average values for each hazard are then added to the output Derived Postcode Database as ‘occurrence factors’.

Figure 1 and Figure 2 show the distribution of the data within the DPD, as represented by the numbers of postcodes plotted against the Total Occurrence Factor (TOF). The total range of the TOF is 0–213. The distribution is quite stepped. This is because the GeoSure hazard rating is broken down into the five classes (A-E). Although the change in hazard between each class is incremental (E is one step worse than D is one step worse than C and so on) the likely occurrence of each level of hazard may vary by an order of magnitude between classes. Figure 1 and Figure 2] also show how natural breaks have been selected using the Jenks’ Classification method in the distribution in order to derive the fivefold (Class 5) and tenfold (Class 10) classifications displayed in the DPD (last two columns of Figure 1]).

These classifications are provided in order to aid interpretation of the hazard scale and also to allow easier representation of the data in mapping tools (e.g. by colouring up the postcode points in a GIS if so wished).

In detail, the Total Occurrence Factor (TOF) class breaks in the DPD in version 7 of the GeoSure Insurance Product are as follows (against the overall 0–213 scale of the TOF):

Class 5 breaks are: 0,14,27,52,115,213
Class 10 breaks are: 0,7,17,24,29,36,52,107,116,133,213
Figure 1 Distribution of postcodes in the DPD against the TOF (0-213) range; also showing position of the Class 5 natural breaks.
Figure 2 Distribution of postcodes in the DPD against the TOF (0-213) range; also showing position of the Class 10 natural breaks.

Interpreting the total occurrence factor scale in the derived postcode database

The results from the DPD have been compared against the geological maps, to make the following broad correlations and recommendations for use of the 0-213 Total Occurrence Factor (TOF) scale.

These recommendations, particularly in terms of predicting insurance risk, should only be treated as a guide. The GeoSure Insurance Product is essentially ‘geological’ in nature, as a predictor of natural ground movement. Any predicted levels of insurance risk due to this ground movement are necessarily qualitative and could only be tested quantitatively against external insurance datasets (e.g. claims data) or the experience of insurance professionals.

TOF values of 95+

Very high TOF values, of greater than c.95 (i.e. upper Class 4 and Class 5 of the fivefold Class 5 division) are almost entirely the result of shrink-swell hazard and/or compressible ground with high GeoSure ratings. This is a reflection of the widespread nature of these hazards and their significance in terms of ground movement. For example, shrink-swell clays with high GeoSure ratings (such as London Clay) are very widespread in SE England.

The very highest TOF ratings result from those areas with a high GeoSure rating for compressible ground AND a high GeoSure rating for shrink-swell. In reality, this occurs where clays with a high shrink-swell potential are overlain by peaty ground, which is a relatively rare geological occurrence (e.g. in North East Anglia). Therefore, TOF values in the 95+ range should be considered a very significant hazard to low-rise buildings and a very significant insurance risk, and usually mean highly shrink-swell clays and/or very compressible (peaty) ground are present.

Reference to the individual Occurrence Factors in the DPD (and comparing the values to their ranges as quoted above) and/or to the HV layers in the GIS should indicate to users which of the hazards are present.

TOF values of 40-94

TOF values within the range 40–94 range (upper Class 3 and lower Class 4 of the fivefold Class 5 division) are relatively rare in the dataset (Figure 1) because of the way the data has been calculated. If values in this range are encountered they may result from the contribution of a single hazard or the combined effects of two or more hazards.

In general terms, TOF values towards the top end of this 40–95 range (i.e. into Class 4) indicate that one or more hazards may be present that are a significant hazard to a low-rise building and of significant insurance risk. Again, in general terms only, TOF values towards the lower end of the range (i.e. into Class 3) indicate that one or more hazards may be present, but that they are only likely to be of possible hazard to a low-rise building and of possible insurance risk, particularly if modifications are made to the ground, for example, changes to groundwater by tree removal, or building work.

Reference to the individual Occurrence Factors in the DPD (and comparing the values to their ranges as quoted above) and/or to the HV layers in the GIS should indicate to users which of the hazards are present.

TOF values of 10 to 39

TOF values in the 10–39 range (higher Class1, Class 2, and lower Class 3 on the fivefold Class 5 scale), which are common in the dataset, indicate some form of natural ground stability hazard being present at a level which may cause ground movement.

The combination of anticipated severity of any hazard event and frequency of the hazard event means, however, that it is only likely to be of possible concern for property insurance. However, in such areas, caution should still be exercised as local variations in the geology, or local effects (for example, changes to groundwater by tree removal) or modifications to the ground, could exacerbate the effects of these hazards such that they may become of concern to property stability.

Reference to the individual Occurrence Factors in the DPD (and comparing the values to their ranges as quoted above) should indicate to users which of the hazards are present.

TOF values of 1–9

TOF values in the 1–9 range (lower Class 1 on the Class 5 scale), which are common in the dataset, indicate the presence of one or more ground movement hazards, but at a level which is likely to be insignificant in terms of property insurance.

TOF values of 0

Zero TOF values indicate that no GeoSure hazards are mapped at that location.

Dataset history

The GeoSure Insurance Product is directly derived from the master GeoSure dataset. The GeoSure dataset is re-processed frequently so as to consider any enhancements in the understanding of the processes that drive the hazards. It is also re-processed in response to better data becoming available; this includes any new or improved geological mapping. This approach enables GeoSure to be based upon the most up to date information available.

The GeoSure Insurance Product is re-issued following updates to the Ordnance Survey Code- Point® Open database.

Coverage

Data is provided for Great Britain. The Isle of Man is not covered.

Figure 3 The coverage of the GeoSure Insurance Product (GIP) dataset.

Data format

The GeoSure Insurance Product (GIP) dataset has been created as point data and are available in a range of GIS formats, including ArcGIS (.shp), ArcInfo Coverages and MapInfo (.tab). More specialised formats may be available but may incur additional processing costs.

Limitations

  • The spatial resolution of the HV layers data is the same as the GeoSure data they are derived from, which, therefore, requires that a minimum search radius of 50m around a site or property be utilized in any application (in addition to any site, property or other search area).
  • Licensing of any GIS software (e.g. ESRI) and topographical information (e.g. Ordnance Survey) used when displaying HV layers is the responsibility of the end users.
  • GeoSure is concerned with potential ground stability related to NATURAL geological conditions only. GeoSure does NOT cover any man-made hazards, such as contaminated land or mining. The only exception to this is the Compressible Ground hazard layer, which does consider man-made ground e.g. landfill.
  • GeoSure is based on, and limited to, an interpretation of the records in the possession of The British Geological Survey at the time the dataset was created.
  • An indication of natural ground instability does not necessarily mean that a location will be affected by ground movement or subsidence. Such an assessment can only be made by inspection of the area by a qualified professional.