to biotic or abiotic conditions resulting from increasing iiuman disturbance, and their biological importance. 

 For each sampling site, the response value for each metric selected is transformed to a metric score. The 

 score for each metric is based on the degree of deviation of the response value from that expected at a similar 

 site under conditions of minimal human disturbance. The individual metric scores are then aggregated to 

 produce an overall indicator score. A higher score indicates better ecological condition {i.e., closer to the 

 expected condition when human disturbance is minimal). More detailed descriptions of the general approach 

 used to develop multimetric indices can be found in Fausch et al. (1984), Karr et al. (1986), Karr (1991), 

 Plafkin etal. (1989), Gibson (1994), Barbour ef a/. (1995), Simon and Lyons (1995), and Karr and Chu (1997). 

 Simon and Lyons (1995) and Karr and Chu (1997) summarize and address criticisms of the multimetric 

 approach to developing indicators of condition. 



Multimetric indicators based on modifications to the original IB! concept of Karr (Simon and Lyons 1995) have 

 been developed for use in various geographic areas in the United States and elsewhere (Miller et al. 1988, 

 Lyons etal. 1995, Yoderand Rankin 1995, Lyons etal. 1996, Hughes and Oberdorf 1999), various systems 

 (Jordan etal. 1993) and taxa (Lenat 1993, Kerans and Karr 1994, Fore etal. 1994, DeShon 1995, Fore etal. 

 1 996, Barbour et al. 1 996). Many of these studies address evaluations of the indicator, approximating some 

 of ORD's evaluation guidelines as outlined in Chapter 1, and use a variety of approaches to address a 

 particular guideline. 



Table 4-1. Rationale for indicators of ecological condition based on stream fish assemblages^ 



Historical data available 



• Autecology of most species described 



• Integrates effects of stressors at various scales, time periods, and 

 levels of organization 



• Includes long-lived and mobile species 



• Assemblage composed of populations, individuals 



• Important resource to humans 



• High level of familiarity with general public 



Compiled and summarized from Karr etal. (1986), Plafkin etal. (1989), Simon (1991), and 

 Simon and Lyons (1995). 



The stream fish assemblage indicator is being developed using data collected as part of the Mid-Atlantic 

 Highlands Assessment (MAHA). This study was funded by the EPA Environmental Monitoring and Assessment 

 Program (EMAP; e.g., Whittier and Paulsen 1992) in conjunction with a Regional-EMAP (R-EMAP) effort 

 (U.S. EPA 1997). The MAHA study represents a partnership between EMAP and EPA Region III to develop 

 and demonstrate EMAP approaches such as probability-based survey designs and appropriate indicators of 

 ecological condition to address specific assessment questions of interest to the Region. 



The monitoring framework for MAHA consists of a regional-scale probability-based survey design to select 



sampling sites. This design permits unbiased inferences to be made with known certainty from the subset of 



sites where samples and data are collected to explicitly defined populations of ecological resource units 



(Larsen 1995, 1997, Diaz-Ramos ef al. 1996). For MAHA, populations are defined based on the total length 



of streams. The design allows one to estimate the total length of streams in the target population (e.g., all 



permanent streams appearing on a particular scale of map) which meet some criteria (e.g., all first-order 



target streams, all target streams within a specific ecoregion). The distribution of indicator scores can then be 



examined for these defined populations to determine the estimated length of stream characterized by a 



particular set of indicator values, with associated uncertainty in these estimates represented by confidence 



bounds. 



4-2 



