‘irregularly flooded’ was interpreted as above Mean Higher Water. Both categories 
were excluded from the frame. 
The study utilized a stratified random sampling design. Within Washington, 
sampling effort was distributed such that Puget Sound received 25 stations, there was 
an intensification of effort in Willapa Bay for a total of 30 stations, and the remaining 
estuaries of the state received 13 stations, for a total of 68 stations. In Oregon, there 
was an intensification of sampling in Coos Bay (30 stations), while 35 stations were 
located within the remaining estuaries of the state, for a total of 65 stations. 
The study design in California was more complex. Two pilot study areas were 
defined, Southern California (Point Conception to the Mexican border) and San 
Francisco Bay (downstream of the delta), each of which had 30 sampling stations. An 
additional 30 sites were randomly allocated along the California coastline outside of 
these intensification areas, for a total of 90 sites. The San Francisco Bay study area 
differed from the remainder of the study by dividing sampling effort approximately 
equally between three habitat types, tide flats, low marsh, and high marsh, with high 
marsh being a habitat type excluded from the remainder of the West Coast intertidal 
sampling frame. For both the San Francisco and Southern California pilot study areas, 
a two stage randomization procedure was followed. As the first stage, wetland systems 
were randomly selected from a list of systems, and then a point sampling location was 
randomly selected within the selected wetland. The advantage of this approach is that it 
allows condition estimates based on percentage of systems, while at the same time also 
allowing for the areal extent estimates of condition being used for all other geographic 
components of the intertidal survey. Wetlands systems are typically managed as 
discrete units rather than as continuous resources, and the two level randomization 
design provides the potential to report on the percentage of wetland systems that are 
meeting their management goals. 
Each sampling region is termed a multidensity category. For each multidensity 
category (see Appendix Table 1), geographic coordinates for the number of primary 
target stations described above were determined during the study design process. 
Additionally, each multidensity category except for the California pilot studies had 
random coordinates for nine times the number of primary stations selected as alternate 
sampling locations. The two California pilot studies had 1.5 times the number of 
primary sites selected to serve as alternate locations. Alternate locations would be 
sampled in the event a primary site was rejected for any reason, such as safety 
concerns or access issues. 
3 
