WP4. Assessment of environmental impacts and resulting externalities from multi-media (air/water/soil) impact

pathwaysA. Rabl, T. Bachmann, R. Torfs - 26 May 2003

Participant ARMINES(WP leader)

USTUTT.IER

VITO

Person-months 11,5 5,8 7

Deliverables Due

D4.1: Inventory of emissions to water and soilfor the major fuel chains.

June 01

D4.2: Report on soil and water pathways,including adaptation of EUSES multi-mediamodel for calculation of site-specific damages.

April 02

D4.3: Results for damage estimates. Dec 02

Pathways taken into account for health impacts of air pollutants.

freshwater

soil

air

agriculturalvegetation

milk meat

saltwater

seafood

ingestiondose

fresh waterfish

deposition (wet & dry)

emission

inhalationdose

Direct emissions to soil or water are a special case where the analysis begins at the respective “soil” and “water” boxes. In the present version seafood is not yet included.

Linearization of the Dose-Response Functions

DRF at two values of the background dose, D1 and D2. For small variations around the background the DRF can be approximated by straight lines with slopes s1 and s2. The

slopes can be different between groups with different backgrounds or different sensitivities. For the calculation of population-total impacts one can take the population-

weighted average of the slopes. Doses below a NOAEL are a special case with slope zero.

00 dose

response

D1 D2

slope s1

slope s2

Justification for Equilibrium Models

One can show thatsince with linear dose-response functions only the collective dose matters for the total impact (irrespective of how it is distributed in

time or among individuals), a dynamic model consisting of compartments with first order processes, yields exactly the same

result as an equilibrium model with the same compartments, regardless of any detail of the time history of the inflow.

Therefore an equilibrium model is sufficient for calculating the total dose, even though the real

environment is never in equilibrium.

Uniform world model (UWM)

For inhalation• verified by comparison with about 100 site-specific EcoSense calculations (EU, Eastern Europe, China, Brasil, Thailand, …);• recommended for typical values for emissions from tall stacks, more than about 50 m (for specific sites the agreement is usually within a factor of two to three; for ground level emissions damage much larger; apply correction factors).

For ingestion it is even better, because food is transported over large distances average over all the areas where the food is produced.

Most policy applications need typical values (people tend to use site specific results as if they were typical precisely wrong rather than approximately right)

Results for DosesCollective doses for central European conditions in mg/yr (for emission 1 kg/yr) = intake fractions x 1E6

by exposure pathway as a percentage of the total (figure) and in mg per emitted kg (table) for base case (tcut = 100 years).

As, Cd, Cr and Ni are modeled as PM10 and Pb as PM2.5; Hg is modeled as metallic Hg for

inhalation, methyl Hg for ingestion. Pathway Arsenic Cadmium Chromiu

mMercury Nickel Lead

Inhalation 3.9 3.9 3.9 22.0 3.9 7.1

Water 15.6 15.8 15.5 2.5 15.8 17.8

Cattle milk 153.5 0.3 37.3 10.6 26.7 10.7

Cattle meat 13.6 1.3 36.1 7.4 43.1 4.1

Freshwater fish 7.8 15.8 0.8 1261.3 15.8 4.4

Grains 60.6 119.3 60.2 217.1 64.2 80.2

Root vegetables 12.4 24.1 12.0 16.3 13.1 15.9

Green vegetables

16.2 47.5 15.9 9.2 17.6 23.9

TOTAL 283 228 182 1547 200 164

Results for Doses, cont’d

Arsenic

Cadmium

Chromium

Mercury

Nickel

Lead

TOTAL (Base case) 283.5 228.0 181.6 1546.5 200.2 164.1

Green vegetables 16.2 47.5 15.9 9.2 17.6 23.9

Root vegetables 12.4 24.1 12.0 16.3 13.1 15.9

Grains 60.6 119.3 60.2 217.1 64.2 80.2

Freshwater fish 7.8 15.8 0.8 1261.3 15.8 4.4

Cattle meat 13.6 1.3 36.1 7.4 43.1 4.1

Cattle milk 153.5 0.3 37.3 10.6 26.7 10.7

Water 15.6 15.8 15.5 2.5 15.8 17.8

Inhalation 3.9 3.9 3.9 22.0 3.9 7.1

Arsenic Cadmium Chromium Mercury Nickel Lead

Comparison with CalTOX (a model based on fugacity)

Ratio of total doses calculated by our model (UWM) and by CalTOX, after multiplying the CalTOX results by the ratio 80/29 of population densities in

central Europa and the USA.

Ratio of doses As Cd Cr Ni Pb

UWM/CalTOX 0.61 0.07 0.39 0.60 0.05

Results for Impacts and Social CostsCRFs, DRFs and impacts, per kg emitted, for the carcinogenic metals.

Unit risk and slope factor from IRIS http://www.epa.gov/iris.

As Cd Cr-VI Ni

Inhalation

unit risk [cancers/(pers·70yr·g/m3)]

4.30E-03 1.80E-03 1.20E-02 2.40E-04

sCR [cancers/(pers·yr·kg/m3)] 6.14E+04 2.57E+04 1.71E+05 3.43E+03

Cancers/kg, inhalation, UWM 2.32E-05 9.73E-06 6.49E-05 1.30E-06

Ingestion

slope factor [cancers/(mg/(kgbody·day))]

1.50E+00

sDR [cancers/kg] 1.07E+00

Cancers/kg, ingestion 3.05E-04

Total cancers/kg 3.28E-04 9.73E-06 6.49E-05 1.30E-06

Cost/kg [€/kg] at 2 M€/cancer 656 19 130 2.6

Damage cost of IQ decrement due to PbDose-response function is quite well determined [meta-analysis by Schwartz 1994]: 0.026 IQ points per 1 µg/l increase of Pb in blood, linear without threshold. Cost 3000 €/IQ point

Two calculations:a) 1.0 g/m3 incremental exposure to Pb in ambient air increases the blood level by 50 µg/l, 1.3 IQ points per child per g/m3

Express in terms of total population (sensitive population 1 to 3 yr)

sCR = 1.43E-2 IQ points/(pers·yr·(g/m3)) 804 €/kg, with relation blood/air

b) relation between blood level Pb and ingestion dose [WHO 1995]

sDR = 3.32E+03 IQpoints/(pers.kg) 1633 €/kg, with relation blood/ingestion. Limit for unleaded gasoline after 2000 is 5mg/l [EC 1998]. At this level, and with 1633 €/kg, the associated damage cost would be 0.008 €/l.

Vito, Rudi TorfsModel comparison

1. Examine Vlier-human (VH) model from Vito• Define most sensitive parameters

2. Set parameters to values used in UWM

3. Calculate (total dose)/(inhalation dose)• To compare with UWM results

4. Do some sensitivity analysis

Example NiWe have made calculations for the following scenarios: 1 VH with the original parameters VH -original 2 VH with the UWM parameters VH+UWM 3 VH with the UWM parameters, but the BCF set to the Flemish values VH+UWM -BCF 4 VH with the UWM parameters, but the soil water distribution coefficient (Kd) set to the

Flemish values, for different pH values VH+UWM -Kd pH = X

Table 1: results for Ni

Ni (kg/(pers.yr)) VH -original VH+UWM VH+UWM -

BCF VH+UWM -Kd pH = 4

VH+UWM -Kd pH = 6

VH+UWM -Kd pH = 8

Soil intake 4.9E-09 - - - - -

Vegetables 1.9E-07 1.5E-08 8.0E-08 1.5E-08 1.5E-08 1.5E-08

Meat 9.6E-09 1.9E-07 3.4E-07 1.9E-07 1.9E-07 1.8E-07

Milk 1.6E-09 8.1E-08 1.4E-07 7.8E-08 7.7E-08 7.7E-08

Water 4.2E-08 3.4E-07 3.4E-07 1.1E-07 3.4E-08 1.1E-08

Dose from ingestion 2.5E-07 6.3E-07 9.1E-07 3.9E-07 3.1E-07 2.9E-07

Dose from inhalation 3.7E-09 3.7E-09 3.7E-09 3.7E-09 3.7E-09 3.7E-09 Total dose/inhalation 68 173 248 107 86 79

(VITO)

Conclusions (VITO)

• Ni : good agreement between the ingestion/inhalation ratio of VH and UWM.

• Cr : mismatch between the two models, perhaps due to the drinking water pathway.

• Pb : VH model with parameters set to the UWM values gives a ratio that is 3 to 4 times higher than that of UWM

• VH-model as a local component?

Conclusions

•Have results for doses and impacts for cancers due to As, Cd, Cr and Ni, and for IQ due to Pb

•Have all the major pathways, except seafood

•Have sensitivity analysis

•What time horizon? Significant contributions >100 yr for Hg and Pb.

•Ingestion dose for toxic metals about two orders of magnitude larger than inhalation dose

•Large uncertainty for doses (order of magnitude?)

•Lack of information on DRFs (for most noncancer effects, e.g. due to Hg, only NOAEL or LOAEL data - not sufficient for quantification of impacts)

•What is included in DRF: only inhalation or also ingestion?