and function as a stationary secondary source of petroleum contamination 

 (Gait 1978). The logical approach to analytical sampling in these situations 

 is to use a sampling grid. Since the areas covered by even small oil spills 

 can be extremely large, the use of a sampling grid for simple random sampling 

 could be both time consuming and costly. Sampling grids, such as the one 

 depicted in Figure 4, are therefore most often used for selective sampling 

 or systematic random sampling of spill areas. 



A recently introduced towed underwater fluorometer (Calder 1978) used for 

 monitoring solublized and dispersed oil in the water column will undoubtedly 

 be useful in delineating future oil spills in the marine environment. Un- 

 fortunately, accurate calibration of this instrument is not possible and 

 hydrocarbon concentrations in the water column are reported in equivalent 

 units of crude oil . 



Shore! ine spills . Spills that occur from a point source on or very near 

 shore should also be sampled so that concentration gradients can, if possible, 

 be defined. This can be accomplished by sampling a series of transects all 

 emanating from the source of the spill and moving outward in a fan-shaped 

 pattern, such as that illustrated in Figure 5A. Another approach to this 

 same problem is to sample a series of transects perpendicular to the shore- 

 line. One transect should run through the source of the spill and the 

 others should be arranged parallel to the center transect. As indicated in 

 Figure 5B, a shoreline impacted over a broad area can also be sampled by a 

 series of parallel transects running perpendicular to the shoreline. Samples 

 should be taken along each transect until background concentrations of the 

 spilled substance are reached. One way to determine when to stop sampling 

 when in the field is to smell the sample and stop when the aroma of the 

 spilled substance can no longer be detected. A sensitive nose can detect 

 gasoline at concentrations of 5 parts per billion (ppb) in water (Melpolder 

 et al . 1953). If this use of the olfactory senses is contemplated, care 

 should be taken to make sure that the spilled substance is not noxious. 



As indicated previously, the minimum number of samples per sampling 

 station should be four. The distance between sampling stations should be 

 determined by the characteristics of each individual site. For example, 

 transects off narrow, steep shores should have sampling stations about e^ery 

 20 feet with the four samples (of sediment) randomly collected at each sta- 

 tion from an area of approximately 10 feet by 10 feet. For long sloping 

 shores, transects should have sampling stations about every 60 feet, prefer- 

 ably with eight samples randomly collected in an area of approximately 30 

 feet by 30 feet. 



SAMPLE COLLECTION METHODS 



A number of specific methods for collecting various types of samples are 

 presented in Table 4. Even though the methods described in this table were 

 used for sampling oil spills, they are also applicable to sampling other 



i 

 187 



