Amplification of Cretaceous Warmth by Biological Cloud Feedbacks

See allHide authors and affiliations

Science  11 Apr 2008:
Vol. 320, Issue 5873, pp. 195
DOI: 10.1126/science.1153883


The extreme warmth of particular intervals of geologic history cannot be simulated with climate models, which are constrained by the geologic proxy record to relatively modest increases in atmospheric carbon dioxide levels. Recent recognition that biological productivity controls the abundance of cloud condensation nuclei (CCN) in the unpolluted atmosphere provides a solution to this problem. Our climate simulations show that reduced biological productivity (low CCN abundance) provides a substantial amplification of CO2-induced warming by reducing cloud lifetimes and reflectivity. If the stress of elevated temperatures did indeed suppress marine and terrestrial ecosystems during these times, this long-standing climate enigma may be solved.

During supergreenhouse intervals of the geologic past, both tropical and polar temperatures were considerably warmer than today, and the gradient between the two was reduced. To even approach these equable climate states with climate models, atmospheric CO2 levels must be specified that significantly exceed most proxy estimates for the Cretaceous and the Eocene (1). Thus, climate modelers have invoked viable but hard-to-evaluate hypotheses of elevated atmospheric methane levels, greater poleward oceanic heat transport, and enhanced polar stratospheric clouds (2).

An unexplored alternative involves planetary albedo, the fraction of the incoming solar radiation that is reflected to space, which is largely dependent on cloud cover and cloud albedo. A major determinant of cloud properties is the abundance of cloud condensation nuclei (CCN). When CCN are abundant, many small cloud droplets form, creating optically dense, high-albedo clouds; when abundance is low, fewer and larger droplets form, creating optically thinner, lower-albedo, and, importantly, shorter-lived clouds (3). Today, pollution dominates continental CCN, producing abundances of thousands per cm3. In remote oceanic regions, biological release of dimethylsulfide is the major pathway for CCN production. Andreae (4) concludes that biological productivity determined the CCN concentrations over prehuman unpolluted land and sea, ranging from a few tens per cm3 in low-productivity regions to a few hundred per cm3 in high-productivity regions, supporting the notion of a prominent role for the biota in climate regulation on the prehuman Earth (5). If CO2-induced warming during supergreenhouse intervals reduced global primary productivity by temperature stress and enhanced vertical stratification of the ocean, causing a reduction in CCN concentration, would lower cloud amounts and albedo have caused further warming?

To explore this hypothesis, we used a global climate model (GENESIS version 3.0) (GCM) to simulate middle Cretaceous [∼100 million years ago (Ma)] climate with various atmospheric CO2 amounts. This GCM has a slab mixed-layer ocean and prognostic cloud water amounts, and version 3 uses the National Center for Atmospheric Research (NCAR) Community Climate Model 3 (CCM3) radiation code with prescribed cloud droplet radii re (3). Cloud droplet radii mainly affect cloud optical depth, infrared emissivity, and precipitation efficiency, Pe, the rate at which cloud water is converted to precipitation. Modern large-scale observations and theory suggest that for ∼10- to 100-fold global reductions in past aerosol and CCN amounts, ∼30% (over ocean) to ∼100% (over land) increases in liquid droplet radii are plausible (3). We simulate the Cretaceous climate with these increases in re and with Pe increased for liquid clouds by a factor of 2.2, reflecting the maximum likely effect of extreme global warmth on marine and terrestrial biological productivity and CCN production rate.

Our Cretaceous model results are shown in Fig. 1. In common with previous GCM studies, increasing CO2 from 1× to 4× preindustrial atmospheric level (PAL) (Fig. 1, A and B) fails to produce the extreme high-latitude warmth implied by temperature proxy data (Fig. 1D). We then performed another 4× PAL simulation with the increases in re and Pe described above (Fig. 1C). Global cloud cover is reduced from 64 to 55%, and the less extensive and optically thinner clouds reduce planetary albedo from 0.30 to 0.24. The ensuing warming is dramatic, both in the tropics and in high latitudes, where it is augmented by surface albedo feedback of almost vanishing snow and sea-ice cover. (Other feedbacks due to changes in cloud types and levels are minor.) High-latitude continental temperatures remain above or very close to freezing year round, in better accord with proxy evidence (Fig. 1D).

Fig. 1.

Annual mean surface-air temperatures (°C) in GCM simulations of Middle Cretaceous (∼100 Ma, low sealevel stand) and zonal averages (A) with CO2 concentration 1× PAL (280 parts per million by volume), (B) with 4× PAL CO2, and (C) with 4× PALCO2 and increased liquid-cloud re and Pe. (D) Zonal average temperatures for land and ocean, land only, and ocean only, with ocean (1) and terrestrial (6, 7) proxy temperature data for the Middle Cretaceous shown as solid rectangles. Dotted line indicates data from simulation with 1× PAL CO2; dashed, with 4× PAL CO2; and solid, with 4× PAL CO2 and increased liquid-cloud re and Pe.

Our results support the hypothesis that widespread increases in re can explain the drastic warming and equable high latitudes during supergreenhouse intervals of the Cretaceous and early Cenozoic. The increases in re could plausibly have been caused by an order of magnitude decrease in CCN concentrations, which we suggest was caused in turn by declines in biological productivity triggered by the climatic consequences of high CO2 levels of ∼4× PAL.

Supporting Online Material

Materials and Methods

Fig. S1



View Abstract

Stay Connected to Science

Navigate This Article