Newfields was requested by Shell to provide a review of the current status of the AgChem Portfolio. The main objectives of the review included the following:

1. To confirm the closure status of the worldwide portfolio, assess the potential for a re-opener and inventory the remaining provision and business risks associated with each site.
2. To identify priorities for short, mid-term and long-term to implement closure of still open sites, identify paths towards closure and eliminate or reduce on going direct costs.
3. To identify opportunities for synergy, consolidation and consistency across the remaining open sites to ensure efficiency in obtaining closure.
4. To develop a detailed scope of work to implement priority recommendations.
5. To identify the viability of reducing/controlling the health coverage and transactional costs associated with Paulinia public civil action.

Since project initiation, NewFields was informed that there are no current provisions for any crop protection site except for the ongoing Ipiranga and Paulinia sites in Brazil. Similarly, it is unclear if any sites are still open given that since the closure of the SC-CPP, responsibility for ongoing actions on sites has been passed to the country chairs/management, and this information is currently not available to NewFields. This information may be available in the SC-CPP data room in London, which we hope to gain access to later. In addition, Shell indicated work on the Paulinia health issues would not be included in the scope given current changes in the site’s management. Therefore, the original scope was modified to account for these, allowing us to focus on identifying the current status of the overall Portfolio including current use of the site and surrounding area, conditions under which sites have been closed and the nature and amount of information available for each site. The review also provides a qualitative ranking of sites based on a high- level assessment of possible residual risks and provides recommendations for follow up.

The Portfolio assessed in this report consists of 88 sites, distributed across 44 countries, that were included in the 2002 and 2004 Newfields Portfolio audits (with the exception of Paulinia and Ipiranga). For the current review, Shell provided Newfields with access to a Shell Sharepoint containing information on a number of AgChem sites. The information available was not consistent for each site and included closed-out reports, EA briefing notes, site summary reports, and in some cases site investigation and human health risk assessments and a few remediation reports.

3.1 Site Screening

To better document and assess the nature of the information available for each site and allowing us to assess the degree of residual risk, a new set of indicators were developed. Non-technical risks were not incorporated as these did not seem to be a valid criterion any longer given the difficulty to obtain such information from the available set of data. Where the location of the site was known, changes to the use of the site and surrounding area since closure was one of the main indicators to assess residual risk, since a change of land use from industrial to residential for example could increase exposure and risk to receptors from potential residual contamination on site, or contaminant impacts off site (i.e. via groundwater migration). The result of screening each site with these indicators has been summarised in a series of graphs to allow a rapid characterisation of the portfolio based on the available information (click to access the Overview Tab).

A quick review of the above data allows the following general observations:

• Out of 88 sites included in the Portfolio, only 31 have their exact location identified. For 11 sites, although the exact location could not be found, a general area for the site was identified which allowed an assessment of their likely current land use. For the remaining sites (46), there was no address or location map available (for 13 of these 46 sites, there was no information at all in the Shell Sharefile; see Table 1).
• The predominantly current use of the sites with location information is industrial.
• Most sites (54), are currently owned by a third party, while 6 or 7 sites appear to be still owned or occupied by a Shell operation. The current ownership status of the remaining 14 sites is unknown. However, it is likely that most of these sites would have been divested a long time ago and therefore owned by a third party.
• 18 sites were closed-out without any intrusive investigation.
• 21 sites have a fenceline community or residential receptor at less than 100 meters.
• Closed-out reports were found for 39 sites.

Table 1. Sites with no information in Shell Sharefile

Site Name	Comments
Jamaica Bell Road
Kenya Nakuru
Kenya Nairobi
Mexico Christianson	Environmental liabilities letter available
New Zealand Woolston
Pakistan Keamari PBS	Confusion between this site and NFU site
Pakistan Vahari
Pakistan Multan
United Kingdom Sittingbourne
Uruguay UTE
Venezuela Merida
Zambia Lusaka
Zimbawe Harare

3.2 General Observations on Remediation Targets

The availability of remediation trigger values was subject to the nature and availability of environmental reports other than EA briefing notes or closed-out reports, as these rarely contain information on risk assessment and remediation criteria. Where available, remediation trigger values varied depending on the exposure scenarios applicable and local regulatory requirements. In some cases, such as former agrochemical sites within active operational sites (i.e. large oil depots or refineries such as SAPREF in South Africa), human health risk trigger values were calculated for both the current situation and a future unrestricted industrial scenario demanding a lower trigger criteria. For example, the unrestricted industrial trigger value for the Reunion site in South Africa was 66 milligrams per kilogram (mg/kg) for aldrin and 74 mg/kg for dieldrin for exposure to surface soils when assessed in 2011, while for the current industrial scenario they were 80 and 210 mg/kg respectively. An unrestricted scenario allowed for changes on the site layout which could increase the exposure to industrial workers (i.e. less surface cover for example).

For other sites (primarily formulation or blending facilities) human health risk assessments were completed assuming a future change of land use to a residential use scenario. Exposure to a child receptor was almost always calculated in these cases. Human health risk assessment for sites in countries with poorly developed contaminated land guidelines followed international best practice (i.e. typically USEPA) with input and approval from environmental regulators. Typical residential (child) values in these cases varied from 0.54 to 2.5 mg/kg for aldrin and 0.57 to 4 mg/kg for dieldrin in surficial soils.

Human health risk-based trigger values for groundwater were less available in the documentation examined, likely due to drins in groundwater being less frequently detected than in soils. Very few risk- based triggered values were identified for groundwater. A Shell screening target level (SSTL) of 0.13 micrograms per liter (ug/l) was mentioned in several reports for dieldrin. Remediation trigger values of 4.4 ug/l for aldrin and 76 ug/l for dieldrin (for construction worker exposure) were derived for the Wadeville site in South Africa.

The current status of the Portfolio can be characterised as having most sites divested and closed out. The responsibility for managing on-going issues in some cases has been passed to the country chair or business management in country. Furthermore, it is uncertain how many sites have not been closed out, as closed-out reports were not located for numerous sites in the Portfolio. It is important to note that most sites were assessed or investigated by Shell between 2003 and 2012 and no major issues appear to have been raised in relation to divested sites since that time, indicating a low likelihood of site re-openers.

Given the current sites’ status, the most likely trigger for future liability will be associated with the following scenarios:

• Change of land use (with or without ownership change) or underground works.
• Property transaction (with or without land use change).
• Impacts to off-site receptors (i.e. off-site migration of contaminated groundwater).

Consequently, given the above triggers, sites included in the Portfolio were placed into the following groups of potential liabilities:

• Sites with suspected or confirmed off-site groundwater contamination.
• Sites with waste/contamination confirmed and with no remediation undertaken.
• Sites that would require further remedial action to meet more stringent remediation criteria should a change of land use occur (i.e. a site remediated for industrial use being divested for residential use).
• Sites with no site investigation (i.e. unconfirmed ground conditions).
• Sites reportedly investigated; however, no data is available.

Liability transfer or restrictions on future use information was limited to six sites, therefore the risk ranking of the Portfolio does not take into account this important aspect for most sites.


Table 2. Liability transfer information

Site	Comments
Nigeria, Lagos Apapa	Liability for remediation passed to Nolchem
Salvador, Usultan	Sold in 1993. Shell retained liability
Iran, Qazum	Sold in 1979. Continued to produce pesticides
Italy, Lana	Sold in 1994. Liability transferred to Lana Municipality
Italy, Mussocco	Sold in 1995 with requirement to continue as industrial activity
Guatemala, Retalhuleo	Liabilities transferred to the purchaser

In addition to assigning sites to the five groups described above, sites within each group were further screened to separate sites into categories of difference qualitative risk, resulting in a different set of recommendations for each.

4.2 Risk Ranking

Ideally, risk ranking would result in some form of risk quantification. Estimating the cost of possible future remedial action is the most obvious way to rank liabilities due to a change in site use, discovery of an unknown condition, and/or redevelopment. However, the data available for most sites did not allow such cost estimations, with only a handful of sites possessing sufficient soil and groundwater data. On the other hand, some sites have sufficient information to confirm that further future actions are unlikely to be needed (i.e. Puente Aranda, Maracay).

Consideration was given to inputting indicators into a mathematical model such as the one used in the 2002 and 2004 NewFields reviews, however the number of indicators available were considered too few for a meaningful result. In addition, many sites could not be located and therefore their current site use and surrounding land use (both important conditions to understand potential residual risks) could not be confirmed. Therefore, sites within each risk category were further screened to assign, qualitatively, different degrees of potential future risk. In this way, sites can be separated into those requiring some further consideration/action and those for which further consideration does not seem necessary. The Ranking Tab in the Explore page shows the five risk categories and subgroups within each category. A brief description of each category is included below.

The risk ranking also includes the likelihood for a site to become encroached by residential occupation in the future, based on the current surrounding land use. A list of these sites is included in the Table 3 below.

Table 3. Sites with possible medium to long-term encroachment of residential habitation

Site	Comments
Paraguay, Asuncion
Philippines, Legaspi
Philippines, Pandacan	Planned commercial/residential development
Uruguay, Bella Vista	Long term possibility
Australia, Newport	Long Term possibility
Chile, Las Salinas	Planned commercial/residential development
Malaysia, Port klang	Site designated industrial/residential. Long term possibility
Morocco, Roche Noire	Long term possibility, apartments likely
Nicaragua, Coquinsa
Nigeria, Port Harcourt
Suriname, Paranaibo
Dominican Republic, Torrecilla	Long term possibility
Philippines, Bacalod	Most likely residential already
Togo, Lome	Long term possibility
Philippines, Davao Sasa

4.3 Group 1: Sites with Known Off-Site Impacts

This category includes sites with known off-site pesticide-impacted groundwater. Three sites were identified with this condition, Australia.Newport, Spain.Maskor and South Africa. Wadeville. Based on groundwater monitoring, modelling, and the nature of receptors, none of the sites were deemed to require remediation. Both Maskor and Wadeville detected drins off-site but their contaminant plumes are under heavily industrialised areas with no groundwater usage. The closest groundwater supply wells are located at distances beyond the extend of the plumes as determined by modelling, and this situation is unlikely to change in the future.

Groundwater immediately off-site the Newport site has reported organochlorine pesticides concentrations under the current residential area. Groundwater was reportedly not used in the area and residential properties were connected to municipal water supply. The site has been given a higher level of potential risk to account for the possibility of a change in the receptor circumstances (i.e. the installation of a groundwater supply well, for domestic use).

4.4 Group 2: Confirmed Soil and Groundwater Contamination

This category includes sites where investigations have confirmed concentrations of pesticides in soil and/or groundwater or where buried wastes were confirmed or are believed to exist on-site based on anecdotal information. However, no information is available as to whether any remediation or removal of wastes was undertaken, or whether remediation was not required based on the site’s particular source-pathway-receptor linkages. Therefore, potential residual risks may exist depending on the future use of the site. Sites were further screened to rank them within risk categories by confirming whether organochlorine pesticides were used instead of other pesticides (higher risk), whether liabilities were transferred to the purchaser (lower risk) and the presence of potential residential receptors (highest risk). Seventeen sites are currently included in this category.

Group 3: Sites Requiring Potential Additional Remediation

This category includes sites that have confirmed soil and/or groundwater pesticides contamination above levels considered safe for the use of the site at the time (typically industrial) and were remediated to safe levels for the continued industrial use of the site (several sites were remediated for unrestricted industrial use). They typically include sites for which Shell did not anticipate a change of use and they represent a potential residual risk if they were to have a more sensitive land use in the future (i.e. residential). Further filtering by ownership identified sites presenting a highest risk (i.e. those owned by third parties), and lower risk (i.e. owned by Shell). There are currently 24 sites in this category.

4.6 Group 4: Unconfirmed Site Conditions

This category includes sites that were closed-out without the undertaking of intrusive investigations. Risk categorization of these sites by the CPP was based on audit reports, anecdotal evidence, site visits, responses from the site, etc, and are indicated generally as presenting a low-risk from the source-pathway-receptor point of view. These types of sites were generally noted to have good housekeeping and operational practices, low risk operations (i.e. storage or packaging rather than liquid formulation), and/or the use of lower risk pesticides (i.e. absence of drins), etc. The decision not to investigate these sites appears to have been reasonable based on their general condition as described above. Further categorization of risk was carried out by applying further filters such as operational type (liquid blending representing a highest risk) and type of pesticides handled (organochlorine pesticides representing highest risk), in an attempt to identify those sites requiring some additional level of attention. Twenty sites are included in this category.

An example of a site on the higher end of potential residual risk in this category is the Retalhuleu site in Guatemala. Although the site was sold with the liability, it presents some aspects that may pose a concern. Although the site has reportedly undergone investigation and remediation there is no actual data available. The exact location of this site in unknown but it is in an area characterised by agricultural activity with small agricultural tenements and residential properties. The plant operations began in 1965 with the blending of liquid insecticides and herbicides, closed in the 1980s and was sold in 1997. It has a long history of on-site incineration and burial of ashes. The site had USTs for xylenes and pesticides (toxaphene and parathion), handled aldrin and dieldrin and up to 8000 drums were stored on-site at any one time. The site is reportedly underlain by clay and gravels and groundwater wells located within 50 m of the site were closed, however, the reason to the well closures is not known. Groundwater in the area was reportedly used for irrigation in the past with potable water supplied from surface water. However, we do not know if this is still the case or if groundwater for other uses currently exist. The site is reportedly located next to a river (River Puca) which could be a receptor from any potential contamination originating from the site. Although the site was closed-out following CPP protocols there are unanswered questions that pose some concern as to possible residual liabilities.

4.7 Group 5: No Remediation Required

This category includes sites that have been investigated as indicated in closed-out reports or briefing notes and have either no impact of soil and groundwater or concentrations of pesticides were found to be below remediation trigger values. Lower risk sites were identified by the actual availability of data while higher risk was assigned to sites were no data was reported and therefore the report relied solely on anecdotal evidence. Higher risk sites also include those with anecdotal evidence of buried wastes. Ten sites are included in this category. Eleven sites are included in this category.

4.8 Sites owned by Shell

The above categories include a few sites that are still owned by Shell. They typically represent former crop protection sites within larger facilities such as refinery (Reunion in South Africa) or large depots (Newport, Australia). A handful of sites that were leased may still be used by Shell. Table 4 includes a list of the sites.

Table 4: Sites owned or possibly owned by Shell

Site	Comments
AUS Newport	Remediated to industrial use standard. Off- site groundwater contamination.
AUS Pikemba	Possibly remediated to residential standard.
DNK Fredericia	No information on pesticides. Significant hydrocarbon impact.
JPN Ichikawa	Seems to be part of the Showa Shell site.
ZAF Reunion	Remediated to industrial use standard.
Leased sites. Unknown if still occupied by Shell
ETH Akaki	Transfer of impacted soils to Europe.
KEN Mumbasa	Safe for industrial use. Groundwater was not investigated.
NZL Seaview	Classified as low risk.
PHL Davao Sasa	Reportedly suitable for residential use.

Of the above list, four sites may require further remediation should they be divested in the future. These are Newport in Australia, Fredericia Harbour in Denmark, Reunion in South Africa and Mombasa in Kenya. Newport, Reunion and Mombasa have had some remediation carried out that is protective for an industrial use. The Fredericia Harbour in Denmark is impacted with hydrocarbons and will require remediation should it be divested in the future. The site has handled crop protection products in the past but anecdotal evidence suggest the site was used for the storage of packed products. We do not know if the site was ever tested for pesticides in soil and groundwater.

With the termination of the SC-CPP Program, we assume responsibility for the future management of soil and groundwater will remain with the site management who will seek advice from PCRO or Shell’s Soil and Groundwater function as required. Any future divestment of these sites will require attention by PCRO to ensure environmental issues/liabilities are effectively managed and minimised.

In the absence of quantification of residual risks, descriptors indicators were used to provide a degree of discrimination between different groups of sites within each of the five Risk Groups. A qualitative risk ranking was completed, based on the selected indicators and professional judgement applied after reading the available reports for each site. Therefore, the discrimination between the High-, Medium- and Low-Risk categories within each Risk Group is not based on any quantifiable measure of risk but is a means of providing a level of management attention for preparedness against potential re-openers.

Since most sites are currently owned by third party companies, we understand that Shell does not intend to intervene on any of these sites based on the findings on this report. Instead, recommendations outlined herein are to ensure that site information is readily available in case of an issue emerging at any of these sites in the future. As such, the prioritization given to the High- and Medium-Risk categories is about filling information gaps such as obtaining missing site locations, missing closed-out reports or remediation verification reports, or obtaining further information from Shell personnel so that risk rankings can be confirmed or adjusted. Available information for AUS Newport, DKN Fredericia, ZAF Reunion and KEN Mombasa should be obtained to allow for the quantification of remediation costs should the sites be divested in the future. We understand that additional information may be held in a data room in the Shell London office and that access to it is currently restricted due to the COVID situation. Table 5 includes a set of recommendations for each risk category.

Table 5. Recommendations

Risk Category
High	AUS Newport: Find recent groundwater monitoring information. Find closed-out report.
	DNK Fredericia Harbour: Find pesticides information the Fredericia Harbour site. Find closed-out report.
	GTM Retalhuleu. Find location and closed-out report.
	MYS Port Klang: Find closed-out report.
	PAK Keamari: Confirm status of solar development covering impacted soils. Obtain closed-out report/liability transfer documents.
Medium	Recover closed-out reports from data room and other sources to complement missing data sources and make them readily available if required in the future.
	Find out liability status of sites.
	Find locations where missing.
Low	Obtain addresses/site locations maps for sites with no location identified.
Additional recommendation	Find information for the 13 sites with no information in Shell Sharefile. Obtain all information for Shell owned sites that may require further remediation if divested.

In the absence of quantification of residual risks, a qualitative risk ranking was completed, based on selected indicators and professional judgement applied after reading the available reports for each site. Most of the sites have been divested years ago and are currently owned by third parties. To date, no issues related to residual contamination or other liabilities have been raised indicating a low likelihood for re-openers. Based on the data reviewed, no major issues of concern have been identified, however, data was not available for many sites to confirm this. Closed out reports are missing for 52 sites and therefore we cannot confirm that the Crop Protection Program process for closing out sites was followed in these cases.

The recommendations given in this report have the objective to raise attention to those sites with a higher risk of residual liabilities and are targeted at finding and collecting relevant missing data for the higher risk sites so that this can be readily available in case is needed in the future. Filling the identified data gaps may also result in changes in the current ranking of some sites.

Angola, Mulemba
Australia, Forest Hills
Australia, Bayswater
Australia, Newport
Australia, Pikenba
Brazil, Londrina
Chile, Las Salinas
Colombia, Palmira
Colombia, Barranquilla
Colombia, Puente Aranda
Denmark, Frederica
Dominican Republic, Coferka
Dominican Republic, Torrecilla
Dominican Republic, Santiago
Egypt, Meco
El Salvador, Usulutan
Ethiopia, Akaki
France, Berre
Ghana, Tema
Ghana, Jamestown
Greece, Sindos
Guatemala, Retalhuleu
India, Lote
Iran, Qazum
Ireland, Seaview
Italy, Lana
Italy, Mussoco
Ivory Coast, Abidjan
Jamaica, Bell Road
Jamaica, Rockford
Japan, Ichikawa
Japan, Kakegawa
Kenya, Nairobi
Kenya, Mombasa
Kenya, Nakuru
Malawi, Blantyre
Malaysia, Port Klang
Mexico, Xalostoc
Mexico, Christianson
Morocco, Roche Noire
New Zealand, Seaview
New Zealand, Woolston
Nicaragua, Coquinsa
Nigeria, Kadura City
Nigeria, Lagos Apapa
Nigeria, Port Harcourt
Pakistan, Badanibagh
Pakistan, Peshwar
Pakistan, Keamari NFU
Pakistan, Suhiwal
Pakistan, Vehari
Pakistan, Keamari PBS
Pakistan, Multan
Pakistan, Hyderabad
Pakistan, Sukkar
Pakistan, Faisalabad
Pakistan, Gujranwala
Panama, Miguelito
Paraguay, Asuncion
Philippines, Pandacan
Philippines, Cagayan
Philippines, Calauan
Philippines, Mactan
Philippines, Legaspi
Philippines, Bacolod
Philippines, Iloilo
Philippines, Davao
Sierra Leone, Freetown
Sierra Leone, Kissy
South Africa, Paarden Eiland
South Africa, Reunion
South Africa, Wadeville
Spain, Maskor
Suriname, Paramaribo
Suriname, Nickerie
Thailand, Bangkok
Togo, Lome
Trinidad and Tobago, Port of Spain
Turkey, Derince
United Kingdom, Sorex
United Kingdom, Sittingbourne
United Kingdom, Widnes
Uruguay, UTE site
Uruguay, Bella Vista
Venezuela, Maracay
Venezuela, Merida
Zambia, Lusaka
Zimbabwe, Harare

Current and future land use
Industrial	1
Commercial	2
Mixed commercial/residential	3
Residential (apartments)	4
Residential (gardens)	5
Agriculture/open ground	6
Unknown	-99

Contamination Status (Risk)
Meets residual child standard	1
Meets Industrial standard	2
Residual contamination above Industrial standard	3
Known off site groundwater contamination	4
Unknown	-99

Contamination Status (Media)
Investigated: no contamination	1
Investigated: soil impact	2
Investigated: soil and groundwater impact	3
Unknown	-99

Current neighbouring land use
No sensitive receptor	1
Env. and/or business receptor	2
Community <100m	3
Fenceline community	4
Unknown	-99

Remediation
No remediation required/carried out	1
Soil and groundwater remediation	2
Soil remediation	3
Removal of buried wastes	4
Unknown	-99

Ownership
Owned/active	1
Owned/inactive	2
Activated/leased	3
3rd party	4
Unknown	-99

Reports
HHRA/remediation/verification reports	1
Closure report	2
Briefing notes	3
No reports	4

Report Data
Soil and groundwater data	1
Soil data	2
No data	3