Quantitative palaeoclimate GCM validation: Late Jurassic and mid-Cretaceous case studies

Journal of the Geological Society, Sep 1997 by Price, Gregory D, Valdes, Paul J, Sellwood, Bruce W

GREGORY D. PRICE1, PAUL J. VALDES2 & BRUCE W. SELLWOOD3

Using quantitative proxy models of peat and bauxite formation, based upon modern analogues, predictions of the distribution of peats (coals) and bauxites for both the Late Jurassic (Kimmeridgian) and midCretaceous (Cenomanian) have been made using a General Circulation Model (GCM). The predictions have been compared with the known distribution of these deposits. Such an approach provides a means of identifying and isolating which components of the GCM accurately reflect palaeoclimate. The prediction of peats and bauxites for the Jurassic and Cretaceous show good correspondence with known coals and bauxites and provides otherwise elusive quantitative estimates of high latitude climate. The GCM is thus successfully replicating precipitation and mean annual temperatures for both the Kimmeridgian and Cenomanian.

Keywords: Jurassic, Cretaceous, palaeoclimatology, bauxite, coal.

Investigations of Mesozoic palaeoclimate using General Circulation Models (GCMs) have often focused on qualitative evaluation of results (e.g. Barron & Washington 1982; Valdes et al. 1995; Chandler et al. 1992). A number of recent studies have, however, highlighted the need for GCM predictions to be quantitatively validated against the sedimentological and palaeobiological record (e.g. Poore & Chandler 1994; Pollard & Schulz 1994; Price et al. 1995). Quantitative estimates of terrestrial climate have been produced by physiognomic analysis of floras (e.g. Herman & Spicer 1996; Wolfe & Upchurch 1987), and have also been derived from organisms such as crocodiles (e.g. Markwick 1994). Comparable results have been used effectively for validating GCM predictions (e.g. Sloan & Barron 1992), but data tend to be restricted to North America and are often Latest Cretaceous, or younger, in age.

The aim of this paper is to compare GCM generated predictions for the distribution of bauxites and peats, using the quantitative proxy-climate models outlined by Price et al. (1997) and Lottes & Ziegler (1994), with the Late Jurassic and mid-Cretaceous bauxite and coal distribution. This will identify and locate the potential areas of bauxite and peat formation in the geological past and permit evaluation of the GCM simulations.

GCM validation. GCM simulations require data both to specify boundary conditions (e.g. palaeogeography, atmospheric CO ^ sub 2^, ocean temperatures) and to evaluate results, but each of these parameters must be kept independent to avoid circularity. Deposits, such as redbeds, evaporites, coals and bauxites, have been used with great success to validate early GCM predictions of palaeoclimate. Inferences from such approaches are often qualitative in character (e.g. 'hot', 'dry', 'humid' and 'wet'), which provide a wide scope for climatic interpretation and may lead to erroneous GCM validation. However, with the proliferation of, and improvement of, GCM studies there has been concurrent demand for more effective GCM validation by quantitative means. This demand has been the impetus for some recent studies which have considered the possibility of using 'conventional' climate proxies such as peat (coals) and bauxites to generate quantitative palaeoclimatic data (e.g. Lottes & Ziegler 1994; Price et al. 1997). These deposits have the advantage of being generally globally distributed and relatively common throughout the latter parts of the Phanerozoic geological record (see Parrish et al. 1982; Chumakov et al. 1995).

The requirements for bauxitization to occur today are, in general, satisfied by warm and humid climates (Bardossy & Aleva 1990; Sellwood & Price 1993). Price et al (1997) detailed comparison of different proxy-climate models of bauxitization against modern bauxite occurrence using ground-based climatic observations as well as GCM predictions for the present. It was determined that climate must exert a primary control on bauxite formation and a proxy-climate model for bauxite formation (based upon that of Bardossy & Aleva 1990) was consequently defined in which mean annual temperature is >22 deg C, precipitation is >1200 mm per year and 6 or fewer months receive less than 60 mm of rainfall.

Coals and lignites have long been taken as indicators of terrestrial humidity. By comparing climate data with a global peat database Lottes & Ziegler (1994) recognized that continuity of rainfall throughout the year is critical for productivity, whilst periods of drought (10 deg C ) strongly curtailed preservation, because falling ground water tables permitted oxidation. These values delineate a broad window for bauxite and peat production to test GCM simulations of the past.

Model description. In order to predict the distribution of bauxites and peats, the proxy-climate models were fed into the UK University Global Atmospheric Modelling Programme (UGAMP) GCM, modified to Late Jurassic and midCretaceous boundary conditions (from A. G. Smith pers. comm.). The version of the climate model used has a horizontal resolution of approximately 4 deg x 4 deg, which allows palaeocoastline data from major continental areas to be accurately delineated. Estimates of orography compare closely with the palaeogeography used in Barron & Moore (1994). Both simulations use a slab ocean, present day orbital parameters and CO2 concentrations are raised to 1080 ppm (4 times preindustrial values), a typical value for the Mesozoic (for details see Valdes et al. 1995, 1996).

Fields which conform with the climatic requirements, outlined above, represent areas where bauxites or peats could form if the climate was the only variable influencing their formation. Fields may occur, however, where no known bauxites or peats exist because they have yet to be discovered or have been removed by erosion or because of unfavourable morphological or hydrogeological conditions. To test the predictions, comparison has been made with Late Jurassic and mid-Cretaceous bauxite and coal distribution, presented in Figs 1 & 2. Evaporite deposits are also shown and generally form in areas which experience warm temperatures, but represent opposite extremes in terms of water balance when compared to bauxites and peat (Lottes & Ziegler 1994; Sellwood & Price 1993). Attempts to model the present-day evaporite distribution, proved unsuccessful, which was primarily related to climate not always being the dominant control upon their distribution (see Sellwood & Price 1993; Price et al. 1995). Theoretically, the formation of evaporites should be mutually exclusive from those areas favouring bauxite and peat generation. However, using whole stages as time frames, results in the geological record being time averaged. Hence, sedimentary sequences, effected by relatively short-term climatic cycles, may be indicative of both arid and humid climates when time averaged (see below).