Impact of Past Global Warming on Biodiversity Atmospheric CO2 (ppm) 6000 Paleosol 4000 2000 Peltasperm −20 −25 Talcher nonmarine organic matter −30 40 30 Age (Ma) 20 10 Figure Cenozoic global change indicators: (a) carbon isotopic composition of benthic foraminifera (Zachos et al., 2001), (b) oxygen isotopic composition of benthic foraminifera (Zachos et al., 2001), (c) mean annual temperature of paleosols in Montana inferred from paleosol salinization index (Sheldon et al., 2002), (d) mean annual precipitation of paleosols in Montana inferred from paleosol chemical index of alteration (Sheldon et al., 2002), (e) mean annual precipitation of paleosols in Utah and Montana inferred from depth to Bk horizon in paleosols (Retallack, 2005b), (f) CO2 concentration in the atmosphere estimated from the stomatal index of fossil Ginkgo leaves (Retallack, 2002) Times of high CO2 are generally warm and wet especially small, porous fossils such as foraminifera Observations of clam or brachiopod shells in petrographic thin sections or the scanning electron microscope for original biogenic microstructure are sufficient to rule out burial alteration (Veizer et al., 2000) A proxy for paleotemperature on land is based on observations that warmer climates have greater evapotranspiration than colder climates, and thus leach more alkali cations from soils The soil salinization index, or molar ratio of potash plus soda to alumina in the clay-enriched subsurface (Bw or Bt) horizons of soils, shows a clear inverse relationship with mean annual temperature in North American soils (Sheldon et al., 2002) Such estimates of paleotemperature from paleosols show trends broadly comparable with marine isotopic records (c) Sydney basin paleosols and glendonites 10 50- 1200 (d) Texas−Oklahoma paleosols 1000 800 600 400 200 End-Permian Mid-Miocene End-Eocene 50 15 - EndGuadalupian 60 End-Paleocene (f) Ginkgo stomatal index EndCisuralian 200 EndCarboniferous 400 Mean annual precipitation (mm) 600 Sovetashan marine carbonate Mean annual temperature (°C) 800 70 Greisbach 8000 (e) Utah−Montana paleosol Bk depth End-Cretaceous Mean annual precipitation (mm) CO2 (ppmV) 10,000 13 C nonmarine organic matter (‰) Mean annual precipitation (mm) 1000 800 600 400 6000 5000 4000 3000 2000 1000 12,000 (b) Carbon isotopes (d) Montana paleosol Bt chemistry 1000 (a) Peltasperm stomatal index and paleosol carbonate 14,000 C marine carbonate (‰) 20 (c) Montana paleosol Bt chemistry 15 10 TRI 13 Mean annual temperature (°C) 18 O marine foram (‰) (b) Marine foraminifera oxygen isotope value Wuchaiping −1 Kungurian Ar tinskian (a) Marine foraminifera carbon isotope value Guada- Loplupian ingian Changhsing Permian Cisuralian Capitanian CA P Q Roadian Wordian Miocene Sakmarian Oligo Asselian Eocene Gzhelian 13 C marine foram (‰) CR Paleo 225 305 295 285 275 265 255 245 Age (Ma) Figure Permian global change indicators: (a) CO2 concentration in the atmosphere estimated from the stomatal index of fossil Lepidopteris leaves (Retallack, 2005c) and carbon isotopic composition of paleosols (Tabor et al., 2004), (b) carbon isotopic composition of marine limestone from Sovetashan, Armenia (Baud et al., 1989) and of organic carbon from Talcher, India (de Wit et al., 2002), (c) paleotemperature indicated by glendonites and paleosols from the Hunter Valley, New South Wales, Australia (Retallack, 2005c), (d) paleoprecipitation from depth to Bk horizons in paleosols of Oklahoma and Texas, USA (Retallack, 2005c) Times of high CO2 are generally warm and wet, and some are also times of significant negative carbon isotope excursions (Figure 1c) This paleosol proxy for paleotemperature can be compromised by burial alteration of soil clay mineral composition, particularly illitization, which can be assessed by X-ray diffractograms of clay (Retallack, 2001) Other indications of paleotemperature on land are not so fully quantified, but nevertheless locally useful (Figure 2c) Deeply weathered kaolinitic soils (Ultisols) are not found at mean annual temperatures lower than 12 1C, whereas soils