ML20151H028

Bldg Inventories:Considerations on Earthquake Potential Losses, Presented at 880412-15 Meeting in Olympia,Wa

ML20151H028
Person / Time
Site:	Trojan File:Portland General Electric icon.png
Issue date:	04/12/1988
From:	Steinbrugge K AFFILIATION NOT ASSIGNED
To:
Shared Package
ML20151H012	List: ML20151H010 ML20151H017 ML20151H028 ML20151H047 ML20151H066 ML20151H071 ML20151H077 ML20151H080 ML20151H093 ML20151H106 ML20151H111 ML20151H116 ML20151H122 ML20151H126 ML20151H138 ML20151H144 ML20151H147 ML20151H159 ML20151H170 ML20151H176 ML20151H185 ML20151H197 ML20151H213 ML20151H217 ML20151H229 ML20151H240 ML20151H253 ML20151H263 ML20151H272 ML20151H275 ML20151P543 ML20151P567 ML20151P703 ML20151P721 ML20151P744 ML20151P748 ML20151P767 ML20151P772 ML20151P775 ML20151P793 ML20151P805 ML20151P816 ML20151P819 ML20151P833 ML20151P842 ML20151P847 ML20151P854 ML20151P858 ML20151P861 ML20151P866
References
NUDOCS 8808010113
Download: ML20151H028 (10)

Text

'~~-~

BUILDING INVENTORIES:

CONSIDERATIONS ON EARTHQUAEE POTENTIAL LOSSES

(

Karl V. Steinbrugge Structural Engineer El Cerrito, California Several disciplines are involved when estimating the amount of damage 1

)

and its monetary loss as a result of earthquake. A few of the necessary geophysical parameters have been covered by Dr. Algermissen. Additionally, the Bruce Olsen panel is scheduled to cover building practices in the Puget Sound I

cities and Portland, Oregon.

Our presentation fits into the foregoing pattern on the simple and prosaic subject of building inventories, but it is is of great importance due to the costs if new ones must be developed. Emphasis is given to the inventory requirements for government vulnerability studies and also those for major financial institutions such as insurance companies, banks, and savings and loans.

' Building inventory

• is broadly defined as a list of all buildings and other structures within a given area which could be damaged in a shock.

Included in the inventory are selected construction, occupancy, function, and value data needed for damage and resulting dollar loss estimates. For certain types of monetary loss estination purposes, damage to tne supporting city

• 11telines' must be included. For example, loss of function and/or loss of I

occupancy due to lifeline failure can be measured in economic teras.

We may sunnarize the driving forces for quantifying inventories from three perspectives:

1. Damageability and its relationships to deaths and injury.

2. Damageability and its relationships to property damage.

3. Damageability resulting in loss of function, including lifelines.

Inventory detail is stallar for casualty and property loss estimates, but it is not identical. 6803010113 880615 PDR ADOCK 05000344 P

PDC

User Needs and Costs

(

For the limited purposes of this paper, the direct monetary loss to large numbers of smaller value properties, such as dwellings, will be given principal attention. However, it is valid to ask why the same iraventories are not used for larger value structures or for individual buildings.

For a single structure, the "real world" hazard analysis may be limited to a proposal which "brings the building up to code" with the implication that life safety is being adequately considered and changing damageability is not a concern. A older flexible steel frame habitational or office occupancy can be strengthened without changing the damageability of high value fragile t

partitions, for example.

Alternatively, it may be partially brought to code requirements in some jurisdictions, or the owner may use insurance as an alternative, thereby commingling life safety with property damage. He may sell the building to pass the hazard to others.

This low cost hazard analysis will cost more than that available for a government vulnerability analysis on a per building basis and yet not be useful for the vulnerability study.

The other end of the cost and hazard analysis scales **a the properties 4

I i

of very large corporations where skilled engineering risk organizations are at i

their best. These risk engineers can and do atudy the construction drawings, inspect buildings and sometimes test the concrete. They any recommend strengthening or demolition, eliminate hazardous features in manufacturing processes, recommend dollar amounts of insurance based on expected loss, and otherwise provide service to the corporation's risk annager. These are very 3

useful reports to all parties. But it appears that more than one major i

corporation may have spent more on their hazard and correction study than did the government on its entire southern California vulnerability study. Both i

aatisfied their user's needs, but their inventories and findings have not been interchangeable.

Financial institutions often obtain these reports end may ask for an independent review. Certain insurance companies which underwrite "highly protected risks" (HFR) have substantial in-house engineering competence. They and a few others can adequately analyze major properties such as aircraft manufacturers. Yalues at risk any represent billions of dollars and their sophisticated all.. hazard approaches are beyond this discussion.

1

(

.a-i

Governmental vulnerability studies for disaster response planners must consider all structures, large and small, in the study area on a uniform basis.

(-

This area any be as large as the Puget Sound region or the Los Angeles basin.

Emphasis is normally on casualties, homeless, and impairment of critical facilities ("lifelines").

My experience as a participant in 10 regional vulnerability studies throughout the western United States shows that gathering building and lifeline inventories any require up to 805 of the available money. Budget limitations do not allow a mathematical at2alysis of individual buildings and only a minimal review of drawings of a very few selected buildings.

l A similar situation exists for the financial institutions. Thousands of homes and businesses any be earthquake insured or have loans against them. Let us examine the economics of hazard analysis on low value properties from an insurance standpoint. For example using expense allocation for dwelling fire insurance policies in California as a guida (Steinbrugge, 1982, Table 8-9), a

$200,000 earthquake insured house would require a $400 annual premium at the usual rate ef $2 per $1000, of which:

Losses, adjustments, production, taxes 86.65, or $347 Internal expense and inspection costs 10.15, or $40 Profit and contingency reserves 3 35, or $13 These dollars are reasonably correct today. Should an underwritor wish to know if a dwelling was superior, average, or deficient before writing the business, he would immediately have to send out an inspector and not wait for a more suspicious time. The costs for even a minimally qualified person would run over a hundred dollars (including car and overhead). This extreze is absurd. Also recall that the insureds must eventually pay for the overheads incurred for those o.rners who were preatua comparing or cidn't buy. More practical alternatives exist, but are difficult to make evat effective.

Quite evidently, loss estimation data aust be gathereu on a simplistic

{

and generalized basis if dwelling earthquake insurance premiums are not to increase.

i

' 1 l

i Vood Frame Dwelline Inventoriest California Examele k

In the interests of time and space, further discussion will be limited to wood frame dwelling inventories in California and their dollar loss estimates. Similar methods apply to other classes of buildings, such as i

tilt-up, unreinforced unit masonry, reinforceo unit ansonry, and the like.

Before an earthquake, non-technical persons such as insurance agents and 4

property owners can provide certain construction related information when the policy is written or renewed. Approximate year of construction, for one example, can be used to good advantage to infer local or regional construction i

practices.

Pre-1933 or post-1933 in California is a general criteria for determining earthquake bracing. Location by city or region along with dwelling age can be related to changes in kinds of foundation, anchor bolting practices, j

veneer anchorage, reinforcement in brick chimneys, improvecents (or otherwise) in wall sheathing practices including its nailing, and numerous other construction features.

Computer prograns using the implications of these data can easily co=pute aggregate regional dollar loss estimates based on previously observed damage patterns and losses. The 1971 San Fernanda and the 1983 Coalinga

(

earthquakes have givet. reliable data for various types of wood frame dwellings (Steinbrugge,1982, Table 6-1 and Figure 6-173 Steinbrugge, et al in press).

The number of insured dwellings, values, ages, and locations can be machine read from internal records. These are then by machine related to the distance from a l

postulated earthquake fault rupture (i.e. region of seismic energy release).

Dollar losses are attenuated for distance. Soil influences on damage are

)

included as well as long period effects.

ZIPS are used as the unit area.

Printouts from these computations provide the aggregate probable maximum loss 1

)

(PHL) for each of an ensemble of earthquakes on selected faults at selected J

epicenters.

Likewise, government economic and vulnerability studies can use f.he same or counterpart data. Similar dollar loan estimates have been prepared with updated California census tract housing and tract value.

l I

-4.

I

Yorification by EarthatGe Eroerienee k

Loss estimates must be based on, or proven by, loss experience.

Post-earthquake field data on over 12,000 dwellings after the 1971 San Fernat4do earthquake are summarized in Table 1.

Table 2 is the analysis summary of the field data in Table 1.

Equation 1 ja a relation between deductible and loss over deductible for values in the second column (All Ages) in Table 2.

(5 loss over deductible) = 8.87 - 0 74 x (5 deductible)

Eq. 1 For geceral loss estimation purposes, this equation is modified for different magnitudes and different soil conditions. Equattor 1 is satisfactory for the of to 105 range, but a more complex equation is desirable for greater deductibles.

Table 3 is a summary of the field survey of every dwelling in Coalinga j

after the 1983 earthquake. Table 4, also of Coalinga, shows one of the comparisons between the agEreEate losses as computed by SteinbruEle et a1 e

(1982) and the losses paid by a major insurance cocpany. The large number of insurance company's paid claims were under a quirk of California law known as concurrent causation, and payments do not include any significant deductibles.

(

The 331 paid losses were compared on a house to house basis, and represents about 15% of the Coalinga dwellings. The correlations are excellent, possibly i

somewhat fortuitous.

One may conclude that appropriately devised post-earthquake field surveys can provide adequate practical loss data for public and private sector use.

I i

]

1 i

Future tvellins Loss Estimation Technieues California orierted loss estination techniques discussed above should be expanded and be kept on a continuing basis, with resse.rch improving these techniques.

In areas such as Salt Lake City, dwelling construction practices have i

also changed over time. Change has taken place from unreinforced brick to wood frame to wood frame with brick veteer. Studies in the midwestern states have shown the need for another set of parameters which often involve the type of

3 -

basement walls and other features. Further work is needed for the Puget Sound

(

and Portland regions.

These kinds of information should be expanded to all seissio areas.

Expansions and improvements should not be made in a vacuus away from the realistic inventory requireper.ts of the public and private sectors.

Finally and certainly not least, compatible post-earthquake field surveys are sandatory for future validation of theoretto studies.

l Cited References Steinbrugge, Karl V. (1982).

"Sarthquakes, Volcanoes, and Tsunamis An Anatomy of Hazards", 392 pp, Skandia America Group, New York.

l Steinbrugge, Karl V., E.J. Fowkes, Henry J. Lagorio, and S.T. Algeraissen (in press).

'The 1983 Coalinga, California, Earthquake Dwelling Monetary Losses" USGS Professional Paper --

I

)

k l

i 4

l l

I I

i 1

i J

i l

4

(

6-1 1

i

i l

l 4

1ARLR 1

('

WOOD FRAME DWILLIBO DAMAQE l

1971 SAN FEREARDO EARTEQDAXE M entane of Buildines Havinn Described Damare Construction Cemeenent None Slimht_

Moderate Severe I

Foundation 91 95 5.85 1.65 0 75 Damage to frame 78.85 16.05 3 35 1 95 i

Interior finish - plaster 4.25 78.45 11.15 6 35 Interior finish - sypsusboard 12.15 78.05 6 55 3.41 i

Exterior finish - stucco (plaster) 20 75 74.15 4.05 1.25 I

' Brick chimney damage 67.65 16.15 6.65 7.45 l

j eTotal brick chimney damage was found in 2 3% of the cases.

I 1

' Total

• means exactly that; essentially no bricks were left i

j standing, or the chinney was otherwise so damaged as to be non-repairable.

i f

j tros Steinbrugge (1982), Table 6-6.

i d

j T

i i

5 i

i I

I i

l l

ll l

I i

l i

}

}

I I

i

(

4 :

.I i

A nauu

(

ynnenn? Lane orna nansetInte yon woon ymanen aver _ rnas 1

3 Percent Loss Over Deductible Values from Field Data 1971 san Fernando, California, Earthquake 1

I

[

j Percent Loma Over Deductible 5

All Dwellinami no Ezeentions All twellinam. Ezeentions Below h All Ames Pre.40 1000-40 Post.40 All Ames Pre-to 1940 40 Post.40 3

i

/

o 9.03 11.84 8.93 8.91 8.16 10.46 7 99 8.13

}

t 1

8.22 11.07 8.12 8.08 7.37 9.71 7.20 7 33 i

j 2

7.41 10 31 7 32 7.26 6.58 8.97 6.41 6.52 3

6.61 9.56 6.52 6.44 5.79 8.23 5.63 5 71 1

4 5.82 8.83 5 73 5.63 5 02 7.51 4.86 4 92 l

5 5 03 8.09 4.94 4.82 4.24 6.79 4.09 4.13 i

i 6

4. 2 's 7 36 4.16 4.03 3 48 6.08 3 33 3 34 1

7 3 53 6.82 3.47 3.28 2.78 5.56 2.66 2.61 l

1 8

2.82 6.28 2.77 2.53 2.09 5.05 2.00 1.87 i

9 2.12 5.76 2.09 1 79 1.41 4.55 1 35 1.14

)

10 1.88 5.44 1.88 1 54 1.21 4.26 1.17 0 92 l

1 t'

L t

i Exceetiens flant 4 eelumns er tableh Dwellings located on faulting or on

)

i

(

observed ground disturbance, split level dwellings,1 and 2 story l

dwellings, and dwellings in areas adjacent to mountains where soil 1

i amplification was observed.

I i

Source: Steinbrugge, ut. published study.

1 1

r i

l i

I l

1 i

i i

i i

l l

1 i

l l( i j

l 4

IABLL1 l

(

voon ranum nunttina ta==as in enitrana As a function of Age and Floor Type a

?

Average i

Number of Forcent i

Dwe111 ann

. tsan I

l j

Fre-1940 age groups i

Wood supported floor 780 28.7 Concrete floor on grede 49 17 9 Seth of the above 799 28.4 l

1940-1949 age group

]

Wood supported floor 287 14.2 t

Concrete floor on grade 34 11.5 Both of the above 321 13 9 i

j l

Post-1940 age group:

Wood supported floor 352 14 3 i

1 Concrete floor on grade 395 9.5

)

Both of the above 747 11.8 1

j All age groups:

{

1 Wood supported floor 1,442 21.2 l

Concrete floor on grade 453 93

(

Both of the above 1.895 18.1 I

j i

{

From Steinbrugge, et al. Table 16. USGS Prof. Paper, j

in press.

l l

l l

3 t

J l

i

[

\\

{

i F

l 1

I f

l t

i l 1

6 1

TABLE 4 COMPARIS3R OF MORETARY La3S ESTIMATES In dollars 1963 Coalinga, California, Earthquake Steinbrugge, et al Insurance Company Ratio of Estimates Paid Claims Columns Are aroue (Column A)

(Column B)

A to B Pre-1940 1,268,000 1,431,000 0.89 1940-49 379,800 317,300 1.20 Post-1949 876,000 553,200 1.58 All ages 2,559,000 2,375,000 1.08 d

From Steinbrugge, et al, Table 15, USGS Prof. Paper, in press.

)

4 i

(-

i 1

i l

l

( ;

l l

I,