77 
SCALING AND INTERPOLATION ISSUES 
The frequency and spatial coverage of water quality sampling will always remain 
lower in relation to the temporal and spatial scales at which estuarine phenomena 
occur. To overcome this reality, researchers must use innovative sample designs and 
statistical methods. Throughout the Chesapeake Bay Program’s tidal data analysis 
and monitoring network design meetings, many of these issues regarding the inter¬ 
pretation of shallow water monitoring data have been raised, but all were not solved. 
The major issues relating to dissolved oxygen are highlighted below. 
Water quality mapping of dissolved oxygen uses measures from a half-meter below 
the surface. Some consider this type of measurement a weakness given that most 
hypoxic events occur in deep-water or deep-channel habitats. The last five years of 
water quality mapping, however, have revealed that hypoxic events can affect surface 
and shallow waters more than initially recognized. Each mapping cruise collects 
calibration samples and water quality depth profiles at five to eight stations per 
segment. In much the same fashion that fixed station profiles are interpolated in 
three-dimensions using the Chesapeake Bay interpolator (see Chapter 2 and 
Appendix D), the surface mapping data could be interpolated along with calibration 
station and mid-channel, fixed-station depth profiles to enhance volumetric estimates 
of dissolved oxygen. Advancements in monitoring attainment technology that enable 
deployment of automated vertical profilers and surface and bottom buoy monitors 
could also support this effort. Overall, the integration of data types such as contin¬ 
uous monitoring, mapping, remote sensing, and fixed-station profiles poses one of 
the greatest challenges in criteria assessment. 
Water quality mapping cruises cannot cover every shallow-water cove and creek in 
a segment, thus presenting a problem for spatial extrapolation of the data. Criteria 
assessment using the CFD method requires the use of an interpolated/extrapolated 
surface from the entire segment and does not allow for exclusion of unsampled areas. 
Almost certainly, many of the areas outside of the sampling boundary have far 
different conditions than those measured in the shallow waters of the main segment. 
These areas represent only a small percentage of each segment, but the question 
remains whether they contain more valuable habitat than the space they occupy on a 
percentage basis. 
Annually, many of the larger fish kills in Chesapeake Bay occur in these small tidal 
creeks and embayments due to anthropogenic influences or natural conditions. Two 
months after torrential rains in June 2006, a Maryland Department of Natural 
Resources aerial photography survey of the state’s Eastern Shore tributaries revealed 
that most small embayments were still clouded by silt and algal blooms to a far 
greater extent than adjacent open waters. To assess conditions adequately in these 
shallow-water tidal creeks and embayments, a probabilistic approach may be needed 
in conjunction with current shallow-water sampling design in which representative 
small tidal creeks and embayments are sampled by the surface mapping and the 
chapter vii 
Shallow-water Monitoring and Application foi Criteria Assessment 
