because it is highly valued both commercially and 

 recrealionally. The least amount of information avail- 

 able and poorest quality of data occur for the spawning, 

 egg, and larval life stages. Except for a few species 

 (e.g., blue crab), very little data has been generated on 

 specific habitat preferences and in situ environmental 

 ranges. This is particularly true for the smaller forage 

 and/or non-commercial fishes and invertebrates. Gear 

 selectivity, inability to correctly identify larval stages, 

 and difficulty of sampling various habitats limits the 

 development and reliability of this information. In 

 addition, life history data are lacking on some of the 

 commercially important sciaenid and pelagic species. 



Data reliability was also based on experimental design 

 and whether the studies were relatively recent. In the 

 case of limited studies, information was occasionally 

 inferred. An opportunity exists to refine the data 

 presented based on additional reviews. 



Given that the amount and quality of available informa- 

 tion vary by species, by life stage, between estuaries, 

 and even within an estuary, considerable scientific 

 judgment is required to derive or infer spatial and 

 temporal distributions from existing data and available 

 literature. Unfortunately, even the most informed 

 judgment is far from perfect due to the complexity of 

 estuarine systems. Consequently, information on the 

 level of certainty associated with each data element 

 must be presented when synthesizing multiple data 

 sets (Table 6). Appendices 2, 3, and 4 provide a 

 complete summary of the personal communications 

 and primary references used so that readers can track 

 and obtain additional information efficiently. 



Analysis of data content and quality. To assess the 

 overall certainty of the ELf\/IR Gulf of t^exico data, 

 mean data reliability was calculated by estuary, salinity 

 zone, species, and life stage. In this analysis, "highly 

 certain" = 3, "moderately certain" = 2, and "reasonable 

 inference" = 1. Mean data reliability was calculated 

 using values for only those species and life stages 

 known to occur within an estuary, i.e., those with a 

 relative abundance of at least "rare" during some part 

 of the year. This was because species and life stages 

 known to be absent were typically scored as highly 

 certain. 



This analysis identified estuaries, species, and life 

 stages that have the most reliable information, and 

 those with the least. This information suggests spe- 

 cies, life stages, and estuaries that could be the focus 

 of research efforts. Future research should include a 

 comprehensive and consistent sampling program to 

 quantify species distributions and abundances within 

 and across estuaries. In addition, life history require- 

 ments need to be determined, especially for those 



species that may not have economic value, but are 

 ecologically important. 



Mean data reliability of fish and invertebrate data 

 ranged from a high of 2.08 for Florida Bay to a low of 

 1 .00 for Brazos River, with an overall average of 1 .86 

 (Figure 5). In general, the reliability scores reflect the 

 amount of fisheries research that has been conducted 

 within an estuary. Reliability scores were especially 

 high for Florida Bay, Tampa Bay, Barataria Bay, and 

 Galveston Bay, all of which are fairty large coastal 

 embayments. They were especially low for the 

 Suwannee and Brazos Rivers, both of which are fairly 

 small tidal rivers. 



When averaged across estuaries and analyzed by 

 salinity zone, data reliability scores were lower in the 

 tidal fresh zone than in the mixing and seawater zones 

 (Figure 6). This may occur because the selected 

 species are primarily estuarine, not freshwater, and 

 may also be indicative of fewer studies of tidal fresh 

 waters. 



When averaged across estuaries and analyzed by 

 species, mean data reliability scores ranged from a 

 high of 2.49 for brown shrimp to a low of 1 .46 for gulf 

 stone crab (Figure 7). Of the invertebrate species, 

 reliability scores were highest for penaeid shrimp and 

 blue crab. They were fairly low for gulf stone crab, 

 spiny lobster, bay squid, and hard clam. Of the fish 

 species, reliability scores were fairly high for gulf and 

 yellowfin menhaden, bay anchovy, pinfish, spotted 

 seatrout, and Atlantic croaker. They were fairly low for 

 bull shark, sheepshead minnow, silversides, and code 

 goby. In general, the reliability scores reflect the 

 amount of fisheries research directed towards different 

 species. Reliabilities were especially high for species 

 with high commercial value (e.g., penaeid shrimp, 

 menhaden) , recreational value (e.g. , spotted seatrout) , 

 or ecological value (e.g., bay anchovy). Reliabilities 

 tended to be lowerfor species with low commercial and 

 recreational value (e.g., bay squid, silversides, sheeps- 

 head minnow, code goby), even though these species 

 are ecologically important and fairly abundant. Low 

 data reliability scores for gulf stone crab may also be 

 because of its relatively recent recognition as a sepa- 

 rate species (Williams and Felder 1986). 



When analyzed by life stage, data for juvenile and adult 

 life stages were most reliable, while data for spawning, 

 larvae, and eggs were less certain (Figure 8). This 

 reflects the number of research studies that have 

 focused on adult and juvenile life stages. Species- 

 specific studies of spawning, eggs, and larvae have not 

 been conducted in most estuaries. Thus, some of the 

 information for these life stages was inferred from life 

 history studies and data from similar estuaries. 



