Ecology of extreme faunal turnover of tropical American scallops

—The marine faunas of tropical America underwent substantial evolutionary turnover in the past 3 to 4 million years in response to changing environmental conditions associated with the rise of the Isthmus of Panama, but the ecological signature of changes within major clades is still poorly understood. Here we analyze the paleoecology of faunal turnover within the family Pectinidae (scallops) over the past 12 Myr. The fossil record for the southwest Caribbean (SWC) is remarkably complete over this interval. Diversity increased from a low of 12 species ca. 10-9 Ma to a maximum of 38 species between 4 and3 Ma and then declined to 22 species today. In contrast, there are large gaps in the record from the tropical eastern Pacific (TEP) and diversity remained low throughout the past 10 Myr. Both origination and extinction rates in the SWC peaked between 4 and 3 Ma, and remained high until 2-1 Ma, resulting in a 95% species level turnover between 3.5 and 2 Ma. The TEP record was too incomplete for meaningful estimates of origination and extinction rates. All living species within the SWC originated within the last 4 Myr, as evidenced by a sudden jump in Lyellian percentages per faunule from nearly zero up to 100% during this same interval. However, faunules with Lyellian percentages near zero occurred until 1.8 Ma, so that geographic distributions were extraordinarily heterogeneous until final extinction occurred. There were also striking differences in comparative diversity and abundance among major ecological groups of scallops. Free-swimming scallops constituted the most diverse guild throughout most of the last 10 Myr in the SWC, and were always moderately to very abundant. Leptopecten and Argopecten were also highly diverse throughout the late Miocene and early Pliocene, but declined to very few species thereafter. In contrast, byssally attaching scallops gradually increased in both diversity and abundance since their first appearance in our samples from 8-9 Ma and are the most diverse group today. Evolutionary turnover of scallops in the SWC was correlated with strong ecological reorganization of benthic communities that occurred in response to declining productivity and increased development of corals reefs. /. Travis Smith. Scripps Institution of Oceanography, University of California, San Diego, La Jolla, California 92093-0244. E-mail: [email protected] Jeremy B. C. Jackson. Scripps Institution of Oceanography, University of California, San Diego, La Jolla, California 92093-0244, and Center for Tropical Paleoecology and Archeology, Smithsonian Tropical Research Institute, Box 2072, Balboa, Republic of Panama Accepted: 23 July 2008 The isolation of the Atlantic from the Pacific Introduction Anderson and Roopnarine 2003), reef corals (Budd and Johnson 1999), and bryozoans (Cheetham and Jackson 1996, 2000, all of Ocean by the Isthmus of Panama ca. 3.5 Ma , . , , ., .. . , , ,. J which exhibit increased evolutionary turnover (Coates et al. 1992, 2004; Coates and Obando ,, _ _ , , ^ , r • • . • , after 3.5 Ma. Kates of origination and extmc1996; Bartoli et al. 2005) was associated with .. ., ,.rr ,, ,. ' tion vary greatly among different taxa, resultlarge changes in oceanographic conditions, ing ^ very ^rong faunal turnover in some high faunal turnover, and significant reorga^axa such as gastropods, reef corals, and erect nization of the benthic communities in the bryozoans but not in other groups such as hiTropical Western Atlantic (TWA) between valves and encrusting bryozoans (Cheetham about 3.5 and 1.5 Ma (Woodring 1966; Jackson and jackson 1996; Budd and Johnson 1999; and Johnson 2000; Todd et al. 2002; O'Dea et Jackson and Johnson 2000; Todd et al. 2002; al. 2007. Early studies were based largely on Johnson et al. 2007). Greater focus on origiextinction of gastropods (Woodring 1966; Vernation also revealed that overall molluscan dimeij 1978; Petuch 1982; Jones and Hasson versity within the southwest Caribbean 1985; Vermeij and Petuch 1986), but more re(SWC) increased during faunal turnover cent analyses have included bivalves (Jackson (Jackson et al. 1993), despite high extinction et al. 1993; Roopnarine 1996; Anderson 2001; rates of several "paciphillic" taxa that were © 2009 The Paleontological Society. All rights reserved. 0094-8373/09/3501-0006/$1.00 78 J. T. SMITH AND J. B. C. JACKSON the primary focus of most earlier analyses (Woodring 1966; Jones and Hasson 1985; Vermeij and Petuch 1986). Some of the most important ecological changes are apparent only through compilations of relative abundance instead of just lists of taxa (Todd et al. 2002). Predatory gastropods and suspension-feeding bivalves declined significantly in abundance, but not in overall diversity while reef dwelling gastropods became more abundant (Todd et al. 2002). In contrast, other ecological groups of mollusks remained relatively unchanged in abundance. These differences between ecological patterns based on relative abundance data versus those derived only from taxonomic lists emphasize the importance of using abundance data in addition to diversity (McKinney et al. 1998; Jackson and Erwin 2006). However, abundance data can be highly affected by sampling biases, so samples must be collected in a standardized and systematic fashion (Jackson et al. 1999; Jackson and Erwin 2006). A major enigmatic feature of the SWC extinction is that much of the peak in rates of faunal turnover, especially for gastropods and reef corals, occurred up to 2 Myr after the final closure of the Isthmus and associated major changes in TWA environments (Budd and Johnson 1999; Jackson and Johnson 2000; O'Dea et al. 2007). High levels of extinction around the end of the Pliocene and early Pleistocene have also been noted in faunas as far north as Florida (Stanley and Campbell 1981; Stanley 1986a; Petuch 1982,1995; Allmon et al. 1993, 1996) and California (Stanley 1986b; Smith and Roy 2006). However, the relation of extinction patterns in these more northern faunas to the formation of the Isthmus is difficult to define, despite the temporal correlation. Synthesizing the previous work of Petuch (1982), Vermeij and Petuch (1986), Jones and Allmon (1995), and Roopnarine (1996), Allmon (2001) postulated that the major environmental cause of extinction in the TWA was a regional decrease in productivity. This hypothesis is supported by major changes in abundance of different trophic groups in the SWC and by evidence for a regional drop in productivity ca. 4-3 Ma (Todd et al. 2002; O'Dea et al. 2007). However, the relationship between events in the SWC and the North American faunas analyzed by Stanley and Campbell (1981), Stanley (1986a) and Allmon et al. (1993, 1996) is still unclear. In this paper, we describe basic patterns of diversity, origination, and extinction, as well as apparent causes of faunal turnover, for species of the highly diverse and abundant bivalve family Pectinidae (scallops) in the SWC and TEP over the past 12 Myr. The family Pectinidae is monophyletic (Waller 1978, 1984, 1991,1993, 2006) yet exhibits a great variety of life habits and ecology that are useful to dissect how changes in the environment have affected the evolutionary expansion or decline of different functional groups. We also examine patterns of origination and extinction at different taxonomic levels as well as the correspondence between diversity and abundance on a sample-by-sample basis. Use of all the available information provides a much more complete and subtle understanding of evolution and environment than that based on taxon counting alone. Analysis of the relationships between environmental change and scallop life-history evolution will be presented in a separate paper. Materials and Methods Quantitative samples of fossil scallops from Panama, Costa Rica, and Ecuador were obtained from 226 bulk samples supplemented by the more than 350 collections of specimens gleaned from outcrops by members of the Panama Paleontology Project (PPP) over the past 20 years. Paleontologists have studied all of the sedimentary basins included here since the 1920s so that we have been able to build on their early collections and stratigraphic framework. Samples come from three regions: (1) the TEP, (2) central and eastern Panama (referred to here as "isthmian"), and (3) the Bocas del Toro and Limon regions of northwestern Panama and southeastern Costa Rica (Fig. 1). The most important collections from the TEP are from the Borbon and Manabi basins in Ecuador (Pilsbry and Olsson 1941; Olsson 1964; Hasson and Fischer 1986; Aalto and Miller 1999; Landini et al. 2002; Cantalamessa et al. 2005), the Nicoya Peninsula in Costa Rica, and the Burica Peninsula in Panama ECOLOGY OF FAUNAL TURNOVER 79 FIGURE 1. Map of sample areas. Boxed areas enclose the generalized sedimentary basins described in the text. A, Manabi. B, Borbon. C, Burica/Nicoya Peninsulas. D, Canal Zone. E, Chucunaque. F, Bocas del Toro. G, Limon. Dredge samples were collected in the region of all five basins sampled for fossils in Panama and Costa Rica, as well as two areas off Nicaragua and Honduras where no fossils were sampled. Abbreviations: GP, Gulf of Panama; GC, Gulf of Chiriqui; SB, western San Bias archipelago; BT, Bocas del Toro province from the Laguna Chiriqui to the Golfo de los Moskitos; LCH, Los Cochinos, Honduras; CMN, Cayos Moskitos, Nicaragua. (Olsson 1942; Coates et al. 1992). Collections from the isthmian region that have both TEP and SWC affinities include the classic formations from the Canal Zone (Woodring 19571982; Collins et al. 1996b; Johnson and Kirby 2006), the north-central Coast of Panama (Coates 1999), and the Darien region of Panama (Coates et al. 2004). The Bocas del Toro and Limon region is the most extensively sampled and includes the Limon Basin in northeastern Costa Rica (Olsson 1922; Collins et al. 1995; McNeill et al. 2000), and the Bocas del Toro Basin in northwestern Panama (Olsson 1922; Collins 1993; Collins et al. 1995; Coates et al. 2003, 2005; Coates 2006). In the analyses that follow, samples from the Isthmian and Bocas del Toro/Limon regions were combined to constitute the SWC regional fauna. Finally, dredge samples of recent scallops were obtained from the TEP and SWC as a baseline for comparison and calculation of Lyellian percentages for samples of fossils (Fig. 1) (Smith et al. 2006). Bulk samples were first sorted to class (e.g., bivalve, gastropod, coral). Scallops were picked and sorted from the bivalve fraction and identified to species following Waller (1969, 1984, 1991, 1993, 2006). Previously undescribed species were identified using open nomenclature pending more thorough taxonomic comparison with samples from outside the regions sampled. Appendix 1 includes a complete list of all of the species found in this study. Individual samples were combined into faunules (Jackson et al. 1999; O'Dea et al. 2007) to obtain a more representative sample of the composition and diversity of the scallop fauna from different places, environments, and ages through time. A faunule represents a group of samples from a single outcrop or closely adjacent exposures that can be assigned with confidence to the same age and environment. Appendix 2 lists all of the faunules along with their age, environmental conditions, and diversity of scallops used in this study. These groupings do not reflect completely equivalent sampling in terms of range of environment or sampling intensity, but they provide larger sample sizes allowing analysis of patterns within time bins. Diversity was calculated as species richness (S) and Shannon-Weiner diversity (H) (Hayek and Buzas 1997; Foggo et al. 2003). Species richness was calculated from all of the different collections combined. H was calculated from only the quantitative data from bulk samples. We calculated stratigraphic ranges for all species collected on the basis of actual occurrence and range-through data and used 1-Myr time bins to assess general trends in diversity. We also calculated extinction and origination rates through time on the basis of numbers of occurrences and per-taxon rates. We have not calculated confidence intervals for first and last occurrences of species because all of the established methods are based only on occurrences (Strauss and Sadler 1989; Marshall 1994, 1997; Wang and Marshall 80 J. T. SMITH AND J. B. C. JACKSON