Technical Comments

Comment on "Otolith δ18O Record of Mid-Holocene Sea Surface Temperatures in Peru"

Science  10 Jan 2003:
Vol. 299, Issue 5604, pp. 203
DOI: 10.1126/science.1076173

Andrus et al. (1) asserted that oxygen isotope profiles from modern and mid-Holocene otoliths of the Peruvian sea catfish Galeichthys peruvianus, and diversity indices derived from midden collections of fish bones, provide evidence for warmer sea surface temperatures (SSTs). They further inferred weaker coastal upwelling off the north-central and northern coast of Peru before 5000 years before present (yr B.P.). We contend that their evidence is flawed and does not support these conclusions.

Our first concern deals with the use of otoliths from G. peruvianus. The common range for this fish is 6°S to 18°30′S; it is not a “tropical” eastern Pacific species and was not prior to the mid-Holocene (2–10). Hence, its presence at the archaeological Siches site (4°30′S) should imply cooler SSTs during the mid-Holocene than today, not SSTs 3° to 4°C warmer, as proposed by Andrus et al. (1). We further contend thatG. peruvianus is a poor choice as a proxy for SSTs. Contrary to (1), populations of G. peruvianus do migrate latitudinally, especially in response to El Niño events (2,11, 12). They range from the coastline to nearly 150 km offshore and are found from the surface to mesopelagic depths (13–16). Spawning individuals might also enter estuaries like other ariids (17). Thus, sea catfish otoliths could record a complex life history that does not provide reliable paleo-SST reconstructions of nearshore surface waters at a given latitude.

We also question the correct identification of otoliths by Andruset al. G. peruvianus was not originally identified at Siches or Ostra (17, 18) and no modern specimens of G. peruvianus were collected near Siches (1). A mid-Holocene lapillus shown in a photomicrograph submitted by Andrus et al. to the World Data Center for Paleoclimatology (19) clearly differs from the modern lapillus shown in figure 2 in (1). The former is from an undetermined ariid species that likely belongs to one of several taxa now inhabiting the Gulf of Guayaquil (20). It has a regularly convex, almost semicircular ventral margin, whereas the latter otolith from G. peruvianus has a flattened ventral margin with a medial swelling.

Our second concern is with the data from Siches. This site is located at 4°30′S (19, 21, 22), not 4°40′S (1) or 4°20′S (17). Small differences in latitude are important at this site because the SST gradient between 4°30′S and 4°S is very steep—typically 2° to 4°C (23). Therefore, it should be no surprise that isotopically derived SSTs at Siches appear warmer than modern SSTs at Paita (5°04′S). The Siches data should be more appropriately compared with modern Talara (4°35′S) SSTs (24). The high proportion (90%) of warm-tropical taxa at the early mid-Holocene Siches site (1) is entirely consistent with the long-standing high diversity of warm-tropical mollusks on modern beaches, late Holocene beach ridges, and late Pleistocene marine terraces near Talara (25–28).

Our third concern is with species lists from the six sites listed in figure 4 of (1) and on the World Data Center site (19). Andrus et al. attributed a decrease in diversity and trophic level indices after 5000 yr B.P. to an increase in anchovy numbers. However, figure 4 of (1) confounds time with geography. Three of the four oldest sites (high diversity, low anchovy numbers) are from northern Peru, where few anchovies are ever found (29). The two youngest sites, both with low diversity indices, are from north-central Peru. One site (Alto Salaverry) has no anchovy bones. Its low indices result from a great abundance (70%) of sciaenids. Anchovies are only present at a single younger site (Pampa de Las Llamas-Moxeke) for reasons that cannot be inferred from the data.

The anomalous site in figure 4 of (1) is Ostra. This site is older than 5000 yr B.P. and situated in north-central Peru, but has high diversity indices comparable to tropical northern sites. Among the distinctive and common taxa from Ostra (19) are sea catfish (Ariidae), bonefish (Albula), and pufferfish (Sphoeroides)—all of which inhabit or favor shallow estuarine or lagoonal environments (5, 17, 30–32), such as the nearby coeval Santa lagoon (33–36). Incidentally, if mid-Holocene G. peruvianus spawned in the Santa lagoon and Santa River estuary, as presumed by Reitz and Sandweiss (17), any conclusion that seasonally elevated δ18O anomalies in Ostra otoliths are largely due to warmer open-ocean summer SSTs (1) must be considered suspect.

In the absence of persuasive evidence (1, 21, 37, 38), we believe it is still premature to draw any conclusions regarding the Holocene oceanographic history of the Peruvian margin (39). Convincing reconstructions for seasonal and interannual SST variations and for upwelling conditions during the Holocene should more appropriately be obtained through detailed measurements of biogenic carbonates from less mobile organisms, or offshore laminated sediments.


Navigate This Article