ppm for short periods of time. Therefore, in terms of acute (immediate) toxic effect, the dispersed oil concen-trations generated by the initial natural dispersion and the dispersant operation are likely to have had only a limited environmental impact, and concentrations rapidly diluted to less than 1ppm. By June 1996 oil concen-trations in the water column had returned to background levels (1-10 ppb) over the affected area.
Therefore, the dispersed oil concentrations (which rapidly diluted to less than 1 ppm)are likely to have had only a limited environmental impact. The SEA EMPRESS Environmental Evaluation Committee (SEEEC) interim report identified lower shore bivalves, and urchins as the main casualties the dispersed oil, with some species washed up in their hundreds (SEEEC 1996).
With commercial fish and shellfish, however, the major issue is tainting. Ministry of Agriculture, Fisheries and Food (MAFF) introduced a fisheries exclusion zone immediately following the spill. Fin fish, however, were found to have little or no contamination, and the ban on salmon and sea trout was lifted on 3 May, and on other fin fish on 21 May (SEEEC 1996) . Shellfish, especially bivalve molluscs, were more hevily contaminated and have recovered more slowly, though it is likely that the ban on these have been lifted progressively.
Therefore, the major environmental and economic impact from the SEA EMPRESS spill has not been from the dispersed oil (accounting for 50% of the oil released) but from the much smaller percentage, 5-9%, of surface oil which resulted in 6,900 oiled birds (66% of which were common scoter) , a heavy impact on the intertidal zone resulting for example in fewer than 20 of the previous 150 rare cushion starfish remaining in West Angle Bay (SEEEC 1996) , and a shoreline response lasting over 12 months.
CLEAN-UP OF THE BULK OIL FROM THE SHORELINE
(15 February to 29 February)
Approximately 10,000-15,000 t of Forties emulsion impacted 200 km of the South Wales coastline by the 29 February. The HFO released (480 t), mainly impacted on the coastline of Milford Haven.
The initial area of oiling from 15 to 21 February was from Skomer Island to St. Govan's Head. From the 22 February the winds changed direction from northerly, which they had been from 19-21 February, to southerly. As a consequence, the remaining surface emulsion which had not evaporated or dispersed, headed towards the shoreline of Carmarthen Bay. Emulsified oil first started coming ashore in Carmarthen Bay on the 24 & 25 February between Pendine and Tenby. The most significant oiling of Carmarthen Bay occurred during the period of the 27 to 29 February.
Beach cleanup operations, which relied heavily on manual techniques and light equipment in order to minimise the impact of the cleanup operation itself, removed the majority of the bulk oil from accessible sites from the 24 to 29 February.
SECONDARY CLEAN-UP AND POLISHING OF IMPACTED SHORELINE
(1 March 1996- May 1997)
By March 1996, the majority of the bulk oil had been removed from the beaches in Milford Haven and Car-marthen Bay. However, there were still several oiled areas of concern including inaccessible coves and sites con-taining contaminated cobble and shingle which were stranded as the high tide receded. Residual oil on beaches where primary cleanup has been completed, still presented a challenge to the response team, as their concen-trations were still of concern in terms of both environmental and amenity considerations. The contaminated areas of shoreline can be categorised into high and low energy environments.
High-energy environment : Surf washing
It had been noted throughout the response that the oil contaminating the shore had not adhered tightly to the coastline. When the samples of oil and beach sediments from Carmarthen Bay were agitated with sea water from the surf zone it was possible, using optical microscopy, to identify a natural interaction between mineral fines and oil droplets. Direct observations under epifluorescence microscopy by Department and Fisheries and Oceans, Canada confirmed that water samples taken from the intertidal zone of Amroth beaches contained oil droplets which were surrounded by the mineral fines as a stable floc (Fig.5 a,b) . The figure shows an image of clay, diatom and oil droplets under phase contrast (Fig. 5a) and UV epifluorescence (Fig. 5b) . Under epifluo-rescence, individual oil droplets are clearly distinguished from the mineral fines. We were thus able to confirm
