Early Warning Strategy - NILU

MA

STER

MA

STER

UCL Presentation for MASTER workshop, January 2006

EUK4

-CT-

2002

-000

93

Early Warning Strategy

Joel TaylorUCL Centre for Sustainable Heritage

Review of Preventive Conservation in EuropeEarly Warning System

Mitigation

MA

STER

MA

STER


EUK4

-CT-

2002

-000

93Review of Preventive

Conservation in Europe

Literature Review – published work on best practice and what should be done

Questionnaire – views on current practice and what is being done

Expert panel – end-user views on current practice and what could be done

MA

STER

MA

STER


EUK4

-CT-

2002

-000

93

Literature Review

MA

STER

MA

STER


EUK4

-CT-

2002

-000

93Gaps in Preventive

Conservation StrategyTechniques often not integrated

Damage functions not known for many materials and difficulty in generalising from one material to a range of materials

Synergy of agents of deterioration

Effects of open display, e.g. historic houses and contemporary exhibitions

Data analysis deficit

MA

STER

MA

STER


EUK4

-CT-

2002

-000

93

(Problems with) Standards

Dosimeters cannot utilise existing standards – why?RH levels often mid-range, and concentrate on physical damage and mould growthTemperature incorporates human comfortLight levels incorporate visibilityFew standards on pollutants (concentration-based)No accepted standards for VOCs

The numbers, rather than the ways of achieving them, gained acceptance.

More recently, standards have evolved from single numbers to ranges

MA

STER

MA

STER


EUK4

-CT-

2002

-000

93Conclusions from the

Literature ReviewPreventive conservation strategies are drawing

upon risk management more and more

Acceptable change is being defined

The rate of deterioration for different locations can be compared, using techniques such as isoperms

An overall strategy can accommodate different aims

Interdisciplinary methodologies are replacing standards

MA

STER

MA

STER


EUK4

-CT-

2002

-000

93

Questionnaire

MA

STER

MA

STER


EUK4

-CT-

2002

-000

93

QuestionnaireQuestionnaire on

monitoring

Sent out to determine how attitudes and practice reflect literature

31 heritage institutions of varying size from 18 European countries responded

Archive Museum Historic House

MA

STER

MA

STER


EUK4

-CT-

2002

-000

93

Perceived causes of damage

02468

101214161820

RH

Tem

p

SO2

NOx

O3Lig

ht/U

VOr

g ac

idsPa

rticle

s

Pests

Hand

ling

Res

pons

esPrimary Cause Secondary causeNot important Unknown

MA

STER

MA

STER


EUK4

-CT-

2002

-000

93

Is monitoring systematic?

Everyone that monitors at all, monitors temperature and RH. Very few people monitor pollutants

For many risks, there appeared to be very little correlation between perceived threat and amount of monitoring -except pests and organic acids (risks with visible impact)

All of the institutions that did not consider temperature or RH as important risk factors (or didn’t know), monitored them

Pollution monitoring is independent of location -proportionally, there is more pollution monitoring in rural locations than urban ones

MA

STER

MA

STER


EUK4

-CT-

2002

-000

93

Questionnaire conclusions

Awareness of risks was related to standards, and therefore intellectual access to information

Pollution and organic acids were not frequently monitored

Rational decision-making is distorted by external factors, such as lack of resources

MA

STER

MA

STER


EUK4

-CT-

2002

-000

93

Expert Panel

First end-user workshop in Krakow

MA

STER

MA

STER


EUK4

-CT-

2002

-000

93

Outcomes of discussion

Light, temperature and RH commonly monitored. Pollutants, less so

Differences in monitoring are more related to the size and type of institution than differences in culture or country

Data overload is the most common problem

Most systems rely on single parameter monitoring, not considering synergistic effects

MA

STER

MA

STER


EUK4

-CT-

2002

-000

93

Overall recommendations

‘Data overload’ a problem

Preventive conservationmethods often not integrated

EWO should help define relationship between risk and damage

EWO should be easy to use

Need for EWO to be relevant to existing methodologies

MA

STER

MA

STER


EUK4

-CT-

2002

-000

93Convergence of findings from literature

review, questionnaire and expert views -1Conservation assessments need to account for

synergy among different risks (all research strands revealed this)

Consistent way of comparing different risks needs to be carried out in practice (all strands)

More data analysis should be encouraged in practice (all stands)

Theory and practice emphasise some risks over others (literature review and questionnaire)

MA

STER

MA

STER


EUK4

-CT-

2002

-000

93Convergence of findings from literature

review, questionnaire and expert views - 2Understanding, and expression, of rate of change

for chemical deterioration is still required (literature review and workshop )

Data overload and lack of integration of methods within preventive conservation (questionnaire and workshop)

Communication between scientists and the general public is required (literature review and workshop)

Integration of existing preventive conservation methodologies is desirable (literature review and workshop)

MA

STER

MA

STER


EUK4

-CT-

2002

-000

93

Early Warning System

MA

STER

MA

STER


EUK4

-CT-

2002

-000

93Developing a new preventive conservation strategy for the

MASTER sensorsIntegrate EWO sensor with existing

preventive conservation methodologies

Understand and express, of rate of change for chemical deterioration

Account for synergy between agents

Encourage more data analysis (simplicity)

MA

STER

MA

STER


EUK4

-CT-

2002

-000

93Existing Preventive

Conservation Methods

Environmental MonitoringCondition AssessmentLocation AssessmentRisk Assessment

Vantaa, 2000

MA

STER

MA

STER


EUK4

-CT-

2002

-000

93FUTURE

PAST

DETERMINISTIC CATASTROPHICDeterioration

Location Assessment

Condition Survey Risk Assessment

Environmental Monitoring

MA

STER

MA

STER


EUK4

-CT-

2002

-000

93

Properties of the EWO-G dosimeter

False PositiveResponse, no damage

False NegativeDamage, no response

True positiveResponse and damage

True negativeNo response or damage

Maritime vapours

Unusual VOCs

Inherent deterioration

Rate of reaction

Physical damage

Biological damage

Dose-response (Tetreault, Brimblecombe, Larsen)

Isoperms (Sebera, Michalski, TWPI)

JNFs (Ashley-Smith . et al.)

Passive layers on objects

Agents reducing reaction rates

Uncertainty (Ashley-Smith)

Stability (Bradley and Thickett)

Deterioration Dosimeter Response

MA

STER

MA

STER


EUK4

-CT-

2002

-000

93Classification in

Preventive Conservation

Building (ASHRAE, 1999; 2003)Condition (Keene, 1991; 2002)Environment (CIE, 1995; Larsen, 1996)Risk (Waller, 1994; 2003)Value (Delta plan, Van der Reyden 1996)

Different kinds of institution will have different aims and requirements in terms of assessment and control

MA

STER

MA

STER


EUK4

-CT-

2002

-000

93

Expected,Acceptable

Very good control

Excellent control

Sensor is not

responding

Sensor is not

responding

External store

Could be better

Expected,Acceptable

Very good control

Excellent control

Sensor is not

responding

Open structure

Poor control

Could be better

Expected,Acceptable

Very good control

Excellent control

Historic housemuseum

Problem with control

Poor control

Could be better

Expected,Acceptable

Very good control

Purpose built museum

Serious problem


Poor control

Could be better

Expected,Acceptable

Archive

54321Calibrated levelsType of

building

MA

STER

MA

STER


EUK4

-CT-

2002

-000

93

Types of institution -11 Archive

Climate controlled and air is purified Possibly more of a storage than exhibition spaceCollections very vulnerable or valuableA realistic optimum

2 Purpose Built MuseumEnvironmental control is possible - Thomson’s Class 1 museum Will probably meet the needs of human comfort and existing standards for organic objectsConsistent with existing ideas of good preservation

3 Historic House museumHistoric buildings and museums with limited resources -Thomson’s Class 2 museum. Most organic objects would be well preserved in this environment

MA

STER

MA

STER


EUK4

-CT-

2002

-000

93

Types of institution - 2

4 Open structureOpen display when the environment is difficult to control May have open windows or little UV protectionFor robust organic objects but not valuable or vulnerable Just outside levels that might be recommended for a museum object

5 External store with no controlVery little control or protection. Environment is open to the outdoors in some respects and provides shelter rather than environmental control. Not to be aspired to but indicates location’s performance

MA

STER

MA

STER


EUK4

-CT-

2002

-000

93Existing Preventive

Conservation Methods

Environmental MonitoringCondition AssessmentLocation AssessmentRisk Assessment

Vantaa, 2000

MA

STER

MA

STER


EUK4

-CT-

2002

-000

93

MA

STER

MA

STER


EUK4

-CT-

2002

-000

93

MA

STER

MA

STER


EUK4

-CT-

2002

-000

93

Strategy

Only required data is analysed

Methods are integrated

Diagnostic

Reduce uncertainty in each method

Does not assume what data exists (or ignore existing data)

MA

STER

MA

STER


EUK4

-CT-

2002

-000

93

Interpreting the dosimeter

1

2

3

4

5Problem may be synergistic, or an unmonitored element of the environment. Assess unmonitored hazards.

Identified hazard causing damage?

Carry out risk assessment, location assessment, monitor environment look at collection.

Once value is assessed, prioritise mitigation.

Responding to identified hazard, possibly a long-standing problem. Diagnostic monitoring recommended.

Location Risk

EnvironmentCondition

Yes

No

Check existing data.

1

2

3

4

5

Expectation

Archive store


Historic house museum

Open display

Outside store with no control

MA

STER

MA

STER


EUK4

-CT-

2002

-000

93

Trøndelag Folk Museum ExampleNew, purpose built museum. HVAC system. Category 2

(purpose built museum gallery)

Environment – ‘Class 1’ museum. Monitor temp, RH and light. No natural light (UV)

Object condition – No visible deteriorationLocation – Checked regularly, no problemsRisk – No risk assessment carried out but major

environmental threats perceived

1121112C

1232112B

O3SO2NO2TWPIUVLightT/RHSite

MA

STER

MA

STER


EUK4

-CT-

2002

-000

93

1

2

3

4

5





Location Risk


1

2

3

4

5

Expectation

No

Yes


Problem may be synergistic, or an unmonitored element of the environment. Assess unmonitored hazards.

MA

STER

MA

STER


EUK4

-CT-

2002

-000

93

1

2

3

4

5





Location Risk


1

2

3

4

5

PPO response (gallery)


Yes

No


MA

STER

MA

STER


EUK4

-CT-

2002

-000

93

Expected,Acceptable

Very good control

Excellent control

Sensor is not

responding

Sensor is not

responding

External store

Could be better

Expected,Acceptable

Very good control

Excellent control

Sensor is not

responding

Open structure

Poor control

Could be better

Expected,Acceptable

Very good control

Excellent control

Historic housemuseum


Poor control

Could be better

Expected,Acceptable

Very good control


Serious problem


Poor control

Could be better

Expected,Acceptable

Archive

54321Calibrated levelsType of

building

MA

STER

MA

STER


EUK4

-CT-

2002

-000

93Disagreement between assessments

Risks are synergistic OR unmonitored ‘type 3’ risk OR present no loss of value

Risk

Problem is with control, not location, OR problem is very subtle

Location

Early warning of future condition by chemical deterioration

Condition

Response on EWO sensor is synergistic, OR a hazard that is not monitored

Environ

EWO-G

RiskLocationConditionEnvironEWO-G

Negative

Positive information

MA

STER

MA

STER


EUK4

-CT-

2002

-000

93Agreement between assessments

\\\ POSITIVE\\\

\\\\\\

NEGATIVE \\\

Risks are unpredictable, OR damage has occurred outside accepted levels

Deterioration is not visible and not expected

Risks are not environmental, or not measured. Check for pests etc.

Risk

\\\POSITIVE

\\\\\\

\\\NEGATIVE \\\

Location is not a cause of deterioration. Inspect data and services

Risk is not environment, or not monitored. Assess risk and condition

Location

\\\ POSITIVE\\\

\\\NEGATIVE \\\

Risk is not environment OR not a monitored hazard. Refer to risk assessment

Condition

\\\ POSITIVE\\\

\\\NEGATIVE \\\

Environ

\\\ POSITIVE\\\

\\\NEGATIVE \\\

EWO-GRiskLocationConditionEnvironEWO-G

Negative

Positive information

MA

STER

MA

STER


EUK4

-CT-

2002

-000

93

Interpretation

Pollutants are the only hazards not monitored. None of the monitored hazards reveal a problem

We can deduce that pollutants are the problem

Pollutant gases are undetected by the museum and the EWO dosimeter reveals the problem

Diagnostic monitoring would reveal that NO2 and O3levels are higher inside than outside (during the field test)

MA

STER

MA

STER


EUK4

-CT-

2002

-000

93Pollutants identified

1

2

3

4

5





Location Risk


1

2

3

4

5


Yes

No


MA

STER

MA

STER


EUK4

-CT-

2002

-000

93

Diagnostic monitoring

EWO-S dosimeter can be used for assessments of external pollution, as a diagnostic monitoring campaign

Shorter exposure time of twenty eight days suitable for investigating problems

Often based on hypothesis testing, monitoring different locations

MA

STER

MA

STER


EUK4

-CT-

2002

-000

93

Mitigation

What to do when a hazard has been determined

MA

STER

MA

STER


EUK4

-CT-

2002

-000

93

Risk chain

release

exposureattack consequence

Risk assessment

Dosimetry Condition assessment

Environmental monitoring

Part of this is determining which points are critical for hazards to have an effect on a collection.

There are various ways of assessing and mitigating damage which relate to different stages of the process.

MA

STER

MA

STER


EUK4

-CT-

2002

-000

93

Dependencies in risk

Outcomes are dependent on a series of prior events, which can be generalised.

What needs to take place for damage to happen?

What are the critical points and pathways in this chain of events?

MA

STER

MA

STER


EUK4

-CT-

2002

-000

93

Dependency modelling

These are deductive, top-down methods of analyzing risks in system design.

It involves specifying a ‘top event’ to analyze (damage).

Followed by identifying all of the associated elements in the system that could cause that ‘top event’ to occur.

MA

STER

MA

STER


EUK4

-CT-

2002

-000

93Damage to object from

pollution

Exposed surfaces for deposition

Interaction of pollutant with

collection

Presence of pollutant in

building

No attractive deposition surfaces

Generation of pollutant internally

Infiltration through natural

ventilation

AND

AND

OR

Infiltration through HVAC

An example…

MA

STER

MA

STER


EUK4

-CT-

2002

-000

93

Relationships in model

In a positively phrased dependency model, ANDdependencies are points of weakness because all events need to occur for the higher event to take place.

OR dependencies represent points of strength, because of alternatives options. Can turn these into ANDs through investigation.

Probabilities can be applied to each event, so cost effectiveness and efficiency can be determined for any action.

MA

STER

MA

STER


EUK4

-CT-

2002

-000

93

Modelling deterministic risk

Events often about extent of impact, not presence or absence of impact. Deterministic risks cannot be modelled this way.

The threshold levels can be used to create ‘steps’ for each event, so pathway has defined levels.

E.g. presence is NO2 at 10 ppb, rather than 5ppb. Steps based on object deterioration and MASTER calibration.

MA

STER

MA

STER


EUK4

-CT-

2002

-000

93

Damage to objects e.g. 5ppb NO2 for one year

Presence of pollutant in

gallery

Presence of pollutant in building

Infiltration from outside

External presence of

pollutant

HVAC Filter

No scavenger in display area

Intake filtration

Intake position

Internal generation

Reaction from NO

Unfluedheating

appliance

Cellulose nitrate

break down

No attractive deposit surfaces

AND

AND

OR

OR

OR

OR

4ppb

6ppb

AND

MA

STER

MA

STER


EUK4

-CT-

2002

-000

93

Acknowledgements

May Cassar and Nigel BladesEuropean Commission 5th FP

MASTER project team

You

Early Warning Strategy - NILU

Documents