Hadley circulation and precipitation changes controlling black shale deposition in the Late Jurassic Boreal Seaway

, Abstract New climate simulations using the HadCM3L model with a paleogeography of the Late Jurassic (155.5Ma) and proxy-data corroborate that warm and wet tropical-like conditions reached as far north as the UK sector of the Jurassic Boreal Seaway (~35°N). This is associated with a northern hemisphere Jurassic Hadley cell and an intensi ﬁ ed subtropical jet which both extend signi ﬁ cantly poleward than in the modern (July – September). Deposition of the Kimmeridge Clay Formation (KCF) occurred in the shallow, storm-dominated, epeiric Boreal Seaway. High-resolution paleo-environmental proxy data from the Kimmeridge Clay Formation (KCF; ~155 – 150Ma), UK, are used to test for the role of tropical atmospheric circulation on meter-scale heterogeneities in black shale deposition. Proxy and model data show that the most organic-rich section ( eudoxus to mid- hudlestoni zones) is characterized by a positive δ 13 C org excursion and up to 37wt % total organic carbon (%TOC). Orbital modulation of organic carbon burial primarily in the long eccentricity power band combined with a clear positive correlation between %TOC carbonate-free and the kaolinite/illite ratio supports peak organic carbon burial under the in ﬂ uence of very humid climate conditions, similar to the modern tropics. This reinterpretation of large-scale climate relationships, supported by independent modeling and geological data, has profound implications for atmospheric circulation patterns and processes affecting marine productivity and organic carbon burial further north


Introduction
Black shales represent major perturbations in the global carbon cycle and are recurrent throughout the Phanerozoic. At the present day, organic carbon (OC)-rich sediments are largely absent from the shallow continental shelves, in marked contrast to the extensive deposits found in epeiric basins and on continental margins and shelves in the past. The challenge for understanding marine black shale distribution, thickness, and-importantly-internal variations in deep time is to better constrain the processes that controlled the location and variability of OC production and burial and their relationship with atmospheric circulation, ocean currents, and dynamic depositional conditions.
There is a general consensus that in shallow marine settings OC accumulated in oxygen-deficient water beneath a stratified water column and resulted from a complex interplay between productivity, preservation, and dilution [e.g., Tyson, 2001]. The relative contribution of each of these factors is debated but each is directly or indirectly linked to atmospheric circulation, through nutrient supply via fluctuations in continental weathering intensity, precipitation and runoff, wind-driven oceanic upwelling, and large-scale surface current systems [Arthur and Sageman, 1994]. A comprehensive assessment of the processes and feedback operating at <10 6 years based on sediment data is difficult and requires both climate simulations and consistent highresolution geologic data from multiple locations [Wagner et al., 2013].
Previous research has proposed that the British sector of the Late Jurassic Boreal Seaway, which connected the Tethys Ocean with the proto-Arctic (Figure 1), was governed by subtropical climate conditions [Sellwood and Valdes, 2008]. Dry subtropical climate is determined by the position of the poleward/descending limb of the atmospheric Hadley cells, which for paleoclimates can only be constrained indirectly in the geological record through precipitation proxies. Further information on the principles of Hadley cell dynamics for the present day can be found in Yin [2005] and for the Mesozoic greenhouse climate in Wagner et al. [2013]. Studies  • Supporting Information S1 • Table S1 Correspondence to: H. A. Armstrong, h.a.armstrong@durham.ac.uk Cretaceous (highlighted below) provide some general insight on the dominant processes and feedback under the paleo-Hadley cells during global greenhouse conditions that may well also have operated in the Jurassic and during other time periods of global warmth: 1. Paleogeography affected the global large-scale atmospheric and marine circulation via modulations (strength and position) of the Hadley-Walker circulation and this affected regional precipitation [Ohba and Ueda, 2010]. 2. Lower latitudinal temperature gradients and poleward expansion of the Hadley cells, with the descending, subtropical limbs located at around 25-30° [Hay et al., 2013]. 3. Large latitudinal net moisture changes associated with an intensification of Hadley cell circulation [Manabe and Bryan, 1985]. 4. A more vigorous terrestrial hydrological cycle leading to an increased nutrient flux to the oceans with implications for marine productivity and enhanced OC burial [Hofmann and Wagner, 2011].
The link between black shale formation and climate is long established [e.g., Jenkyns, 1980], and numerous studies have refined these connections and feedback and their biogeochemical consequences, particularly in open ocean basins [e.g., Wagner et al., 2013] and within the context of short-term global warming [Jenkyns, 2003]. It has been shown that seasonal to orbital-scale fluctuations in runoff, upwelling, productivity, and seawater redox can translate into black shales as millimeter-to meter-scale variations in OC content and quality [e.g., Kuhnt et al., 2005;Wagner et al., 2013], often linked with fluctuations in grain size and mineralogy [Berger et al., 1984]. These meter and submeter heterogeneities in shale can therefore be used to trace climate patterns during times of deposition and test for the links between orbitally modulated climate and OC deposition in the past.  [Sellwood and Valdes, 2008], which we revise in this paper.

10.1002/2015PA002911
Following the conceptual model developed for the subtropical-tropical Cretaceous Atlantic [Wagner et al., 2013], the strongest contrasts in depositional conditions and geochemical properties, driven by orbitally paced wet/dry climate variations, occur beneath the ascending and descending limbs of the Hadley cell. Beneath the ascending limb, the forcing of nutrient flux via upwelling and monsoonal continental runoff is strongest, producing cycles of highly variable organic matter quantity and quality in the sedimentary record. Beneath the descending limb the influence of trade wind forcing produces continuous and generally enhanced OM quantity and quality [Wagner et al., 2013]. The specific response of the Jurassic Hadley cell circulation to variable intensities and frequencies of orbital forcing and its impacts on deposition in epeiric basins, particularly at the regional scale, has not yet been tested on core material, defining the scope of this study on the Kimmeridge Clay Formation (KCF), primarily by using high-resolution data from sections in Dorset and Yorkshire, UK.
Our new climate simulations using the HadCM3L model with a paleogeography of the Late Jurassic (155.5 Ma [Getech, 2013]) indicate warm and wet tropical-like conditions between 35°N and 54°N during boreal summer (June-September) in the British sector of the KCF, previously described as "subtropical" [Sellwood and Valdes, 2008]. Our model is consistent with published high-resolution climate proxy data from the British sector of the KCF [Desprairies et al., 1995;Morgans-Bell et al., 2001;Hesselbo et al., 2009;Huang et al., 2010] that confirm a low-latitude climate control on OC productivity and deposition. We therefore expand the Cretaceous concept that at orbital time scales black shale deposition was directly linked to variation in rainfall intensity associated with the Hadley circulation [Hofmann and Wagner, 2011;Wagner et al., 2013] to the Boreal Seaway of the Late Jurassic. If confirmed, this has fundamental implications for the climate and depositional controls and OC burial further north, into the proto-Arctic.

Geological Setting
Deposition of the KCF occurred during overall global greenhouse conditions with pCO 2 values at least 4 times higher than present atmospheric levels [Sellwood and Valdes, 2008]. Consistent with this global climate state, there is no direct geologic evidence for polar ice sheets at that time . There is a consensus that the KCF in the Wessex Basin of the UK was deposited in a shelf environment below fair-weather wave base but close to storm wave base [Macquaker and Gawthorpe, 1993]. Increasing organic (OC) content of the shales has been used as evidence for transgression during the lower part of the formation [see Morgans-Bell et al., 2001]. Fluctuations in water column stratification and bottom water redox are indicated by sedimentology, ichnology and paleontology [Wignall, 1989;Oschmann, 1991], bulk organic and inorganic geochemistry [e.g., Tyson et al., 1979], and biomarkers [Saelen et al., 2000]. These independent lines of evidence support highly variable seawater redox conditions, from fully oxic to anoxic and euxinic. The presence of isorenieratene and its derivatives in samples from the wheatleyensis to pectinatus biozones indicates that the base of the photic zone was periodically euxinic [Van Kaam-Peters et al., 1998].

Methods
The model simulations are run with the HadCM3L model, using the same configuration and spin-up procedure as in Lunt et al. [2016]. The model is run for a total of 1422 years, with a CO 2 concentration of 1120 ppmv, and a paleogeography of the Late Jurassic (155.5 Ma) including a sea level highstand line and topography defined using the methods of Markwick and Valdes [2004]. The final 30 years of the simulation are averaged to provide the climatologies. From these climatologies, we use an automated procedure to identify the Intertropical Convergence Zone (ITCZ), adapted from Berry and Reeder [2014]. Two methods of identification are used; the first is based on the maximum tropical precipitation (blue lines in Figure S2 in the supporting information), which identifies the ITCZ from its surface precipitation expression, and the second on the maximum midtropospheric (500 mbar) vertical velocity (red lines in Figure S2), which is a more dynamical definition based on the rapid ascent of buoyant air masses. We also identify subtropical zones using similar precipitation and vertical velocity metrics (black and orange lines in Figure S2). The automatically located ITCZ and subtropics for the Late Jurassic using the same identification method applied to modern conditions are shown in Figure S1. Based on these, and the precipitation maps themselves, we define the location of the ITCZ for the Late Jurassic in Figure 1 (red line).

10.1002/2015PA002911
We also identify global monsoonal regions during the Late Jurassic ( Figure 3a) defined using the criteria of Wang et al. [2011], which uses the local summer minus winter precipitation rate that exceeds 2 mm/d with local summer precipitation also exceeding 55% of the annual total. Furthermore, we diagnose the modeled Jurassic large-scale atmospheric circulation, including Hadley cells, in terms of cross sections of the vertical and zonal (westerly) wind speeds (Figures 3b and 3c).
It is important to evaluate how well the model simulates the modern atmospheric circulation relative to observations. Our HadCM3L preindustrial simulation reproduces ( Figure  Linear sedimentation rates (LSR) and mass accumulation rates (MAR) of OC and bulk shale were calculated for each individual sedimentary cycle as defined by Kuhnt et al. [2005], assuming that (1) LSRs were linear within each sedimentary cycle and that (2) each of them represents one long eccentricity cycle of 400 kyr ( Figure 4). The first assumption may well bear some inaccuracy in respect of short-term fluctuations in sedimentation and winnowing; however, we have no further evidence to improve this generalization. The second assumption is based on spectral analysis of % total organic carbon (TOC) [Huang et al., 2010]. To calculate bulk MARs dry bulk density (DBD) was used following MAR bulk = LSR*DBD with DBD obtained from the literature Kuhnt et al., 2005]. %TOC values are corrected for carbonate content using the following: TOC carbonate-free % = [100/(100-carbonate %)] * %TOC. The relationship between kaolinite/illite ratio and %TOC carbonate-free is based on data from Dorset [www.earth.ox.ac.uk/~rgge/data.html]. Figure 2 and Figures S1 and S2 show the ITCZ as identified using the two criteria defined in the methods section. Contrary to the modern situation (for reference also shown in the supporting information), in June/July/August the ITCZ splits as it reaches the American continent from the east Pacific, resulting in a northern arm, which is pinned by the proto-Appalachian mountain range, and a southern arm, which is pinned by the North African mountain range. Of these, the northern arm is strongest and extends to about 30°N in the region of the KCF of the UK. The ITCZ is typically associated with the ascending branch of the Hadley circulation, found where vertical velocities are strongly negative from the surface to the top of the troposphere (units of Pa/s; blue regions in Figures 3b, 3e, and 3h) in the tropics. Typically, the stronger the vertical velocities through the atmosphere, the stronger the precipitation, because air converges at the surface in these regions, entraining moisture, and is then transported vertically. As the moist air is forced to rise, the moisture entrained condenses and precipitates out. shown to be monsoonal ( Figure 3a) we surmise that both a northward extension of the ITCZ and monsoonal conditions in the model will have had a strong influence on the Jurassic depositional environment.

Climate Simulations
Moving eastward, the ITCZ is then deflected southward into the tropical Tethys Ocean. Notably, the area of enhanced humidity is also pulled north on both sides of the NW European Boreal Seaway up to the Arctic Ocean, driven by elevated regions along the coastlines of Canada and Scandinavia. The character of the Indian and East Asian monsoons in the modeled Kimmeridgian had less continental precipitation compared to the modern. As for the modern, in December/January/February, the ITCZ is in the Southern Hemisphere, but the Jurassic ITCZ does not deflect as strongly over South America or Africa. Figure 2 and Figures S1 and S2 also show the location of the dry subtropical-like regions between 45 and 60°N, with the associated descending arm of the Hadley cell shown by the green line in Figure 2. This broadly parallels the ITCZ in the Northern Hemisphere. It extends across Laurentia into the paleo-Arctic Ocean in the vicinity of the Boreal Seaway and southward into central Europe-Asia. At least between 5.625 and 24.375°E longitudes and consistent with the subtropical jet core being centered around 55-60°N (Figure 3c).  [Getech, 2013]). Upper map is boreal summer (June, July, and August), and lower map is boreal winter (December, January, and February). During boreal summer the ITCZ lies in a more northerly location than at the present day placing the boreal sector of the NW European seaway temporarily under tropical-like conditions. The green line shows the approximate 1 mm/d precipitation contour, whose position is influenced by the location of the descending limb of the Hadley cell and the subtropical jet, as indicated in Figures 3b and 3c.

Sedimentology, Depositional Environment, and Astrochronolgy
Much of the Kimmeridge Clay Formation has a high OC content, but the most enriched interval occurs in the wheatleyensis-hudlestoni zones, where TOC frequently exceeds 10% [Tyson, 1996] (Figure 4). The depositional and preservational conditions during these specific periods are therefore of particular interest.
The three lithologies common to much of the formation-grey shale, black shale, and coccolith limestoneare present in the wheatleyensis-hudlestoni zones, suggesting that the depositional conditions were broadly similar to those interpreted elsewhere in the succession. The grey shales of the upper wheatleyensis zone are homogenized. This homogenization could have resulted from physical or biogenic (bioturbation) processes or a combination of the two. In the Clavell's Hard Stone Band, for example, occasional Planolites trace fossils are seen in homogeneous grey mudstone (Figure 5e), providing evidence that the sediment was being intensely processed by infauna.

10.1002/2015PA002911
The succession above the Clavell's Hard Stone Band becomes increasingly OC-rich, with a transition from grey shales to extremely OC-rich black shales (>20% TOC) just below the Blackstone Band. The sedimentology is distinctive, with the black shales being formed of thin, event beds, with a somewhat "lenticular" fabric (Figures 5a and 5d). These lenses might be compressed ripples, bedding-parallel trace fossils, or ripped-up clasts of (microbially bound?) mud, possibly both. Burrow diameters (often >5 mm) indicates that energetic and probably oxic conditions were present at the sediment-water interface; silty, bioclastic mudstones ( Figure 5b) and low-angle ripple cross-bedding (Figure 5i) also provide support for dynamic seafloor processes.
Above the Blackstone Band the %TOC remains high, but the lithology changes again, with the lower hudlestoni zone being marked by increasing quantities of primary carbonate from coccoliths ( Figure 5c). The Rope Lake Head Stone Band and White Stone Band are coccolith limestones (Figures 5f, 5g, and 5h), rather than the dolomitic stone bands of the type present lower in the succession. The Rope Lake Head is also noticeable in containing the first large, ichnofabric-forming trace fossils: a low diversity ichnofauna dominated by faecal pellet-rich specimens of Rhizocorallium (Figure 5f). Burrows of this size (shaft diameter >10 mm) strongly support oxic conditions at the sediment-water interface, at least periodically, and are also seen in the White Stone Band (Figure 5h). The low diversity of trace makers, however, suggests that other restricting factors were present. The occurrence of grey shales homogenized by small infauna, black shales deposited as event beds, and OC-rich coccolith limestones colonized by specialized deposit-feeders are indicative of a more dynamic dysoxic-oxic environment. Our observations are consistent with those of Wignall [1991] who reported the upper dysaerobic biofacies contained a moderately diverse, untiered assemblage with Rhizocorallium and Chrondrites, developed in a surface mixed layer. Figure 6 shows a stratigraphical crossplot between "basin-centered" sections in Dorset and Yorkshire and the more proximal, shoreface section at Boulonnais, northern France. Notably, high-resolution correlations of individual %TOC peaks can confidently be made between Dorset and Yorkshire, over a distance of 400 km; they cannot be made to the Boulonnais section in northern France.
A multiproxy record with a pronounced cyclicity has been obtained from the Swanworth and Metherhills cores from the Wessex Basin [Weedon et al., 2004]. These records were further refined and tuned to the 405 kyr and 100 kyr eccentricity signals [Huang et al., 2010]. Huang et al.

Linear Sedimentation Rate
On average, LSRs are high, on the order of about 6 cm/kyr to 11 cm/kyr (Figure 4). The general pattern indicates the presence of a lower but irregular frequency rhythm being superimposed on the well-developed bed-set cycles. This rhythm is independent of the other parameters and therefore attributable to local factors including topography, bottom water currents, storm activity, and sediment grain size.

Mass Accumulation Rates
Organic matter in the KCF is dominated by Type II marine amorphous kerogen with minor Types I and III [Farrimond et al., 1984;Scotchman, 1991]. Type III kerogen, though generally rare, is most abundant in basin margin settings, while Type II kerogens were restricted to basin centres. Hydrogen indices vary between~120 and 550 mgHC/gTOC [Tyson, 2004] and are lower in proximity to the inferred paleo-shoreline and higher in more distal settings [Scotchman, 1991]. The section containing the highest %TOC is characterized by a positive δ 13 C org excursion that spans the eudoxus to mid-hudlestoni zones. This interval has an average >5 wt % TOC [Morgans-Bell et al., 2001] (Figure 4).
OC burial and %TOC carbonate-free are both high for the Kimmeridge section and show similar patterns (Figure 4). Both increase from low values from the base of the section up to the middle of the hudlestoni zone. They fall back within the lower hudlestoni zone but then rise through the remaining hudlestoni to mid-pectinatus zone. Above the mid-pectinatus zone there is a gradual decline back to low values at the top of the section. Superimposed on the long-term trend are distinct higher-frequency and high-amplitude fluctuations, both parameters ( Figure 4). Fluctuations in TOC are observed on~40 m to submillimeter long cycle scales. These amplitudes are generally higher from the eudoxus to mid-hudlestoni zones.
For much of the record, periods of maximum OC burial and kaolinite/illite ratio are structured into bundles of four peaks per 400 kyr cycle (Figure 7). Based on the age model for the KCF this suggests an assigned frequency of 100 kyr for each of these peaks, and therefore, these likely reflect short eccentricity fluctuations

Continental Climate
In addition to the orbital-scale fluctuations there are clear longer-term trends in the kaolinite/illite record from the Dorset section [Hesselbo et al., 2009] (Figures 4 and 7). The record shows an increase in the ratio from 1 in the cymodoce zone to a peak of 2.6 in the mid-autissiodorensis zone. This is followed by a decline back to a value of 1 at the top of the hudlestoni zone and a further rise to 2 in the rotunda zone.

Discussion
Our new climate simulations indicate a poleward shift to around~35-40°N of high precipitation tropical-like regions, associated with changes in the Hadley circulation during the Late Jurassic greenhouse, likely driven by the clustered distribution of landmasses surrounding the proto North Atlantic and western Tethys region and topographic effects from mountain ranges along the coastlines of North America and North Africa, and along the Boreal Seaway. This climate scenario would be consistent with storm-influenced sedimentation via intensification of tropical storm tracks in a shallow water setting. Sedimentary textures indeed indicate that storms had a significant influence on sedimentation during the most OC-rich parts of the KCF. This is consistent with a variety of storm-produced event beds described from the KCF, including horizons of graded rip-up clasts, silt laminae, thin graded mud horizons, and shell pavements [Wignall, 1989]. A mechanism of storm-induced benthic oxygenation and temperature-stratified inhibition of storm mixing has been proposed for the abrupt nature of the decimeter-scale organic-rich and organic-lean shales in the KCF [Wignall, 1989].
The kaolinite/illite ratio is a well-established proxy for continental climate and associated land surface processes. Kaolinite commonly forms in soils under tropical humid conditions [Thiry, 2000]. In shelf basins, where riverine input of clay minerals is highest and often rapid, lateral transport of the mineral load occurs primarily with strong surface currents and during storm events, preserving an almost direct record of continental climate [Singer, 1984]. An alternative, diagenetic origin for the kaolinite in the KCF has been rejected based on petrography and regional considerations [Hesselbo et al., 2009]. Burial maturation in the Kimmeridge Clay is also compatible with the occurrence of smectite, which is only rarely deposited in the Wessex and Cleveland basins [Hesselbo et al., 2009]. Figure 7 shows the relationship between kaolinite/illite ratio and %TOC carbonate-free records as indicators of wet/dry conditions and OC burial respectively. Superimposed on the long-term (million year) trend are distinct higher-frequency and high-amplitude fluctuations at 40 m and~10 m scales. These high-amplitude and short-term variations occurred synchronously for TOC carbonate-free (and %TOC) and kaolinite/illite ratio on a sample-by-sample basis. Based on a mean LSR of 8.25 cm/ka, these equate with the 405 kyr and 100 kyr eccentricity cycles.
The consistency of this variability in OC MAR and kaolinite/illite ratio in these orbital power bands strongly supports a close link to climate forcing. Sedimentation, as indicated by bulk accumulation and carbonate accumulation rates in the British sector during the KCF, however, did not react in a strictly uniform manner to these orbital climate variations, which we attribute to local effects of topography, currents, sediment grain size, and/or productivity.
%TOC peak to peak correlation between the Swanworth core in Dorset and the Ebberston 87 core in Yorkshire indicate this situation extended over a distance of 400 km. The principle mechanisms responsible for these distinct patterns are hypothesized to be variations in marine productivity followed by stratification-driven anoxia. Furthermore, OC burial and black shale formation must have been intimately linked to climate-modulated fluctuations in nutrient cycling. The supply of clay minerals from continental sources suggests continental runoff as a primary mechanism controlling nutrient supply, and thus black shale formation, consistent with organo-mineral studies of Cretaceous open marine oceanic anoxic event (OAE) black shale sections [Kennedy and Wagner, 2011;Loehr and Kennedy, 2014].
This conclusion is expected from the geological context, as the close vicinity of the Boreal Seaway to the Euro-American continent and the warm global climate during the late-Jurassic should have stimulated an intensified hydrological cycle with high precipitation and runoff [Sellwood and Valdes, 2008].
Orbital-scale variations in OC content are well known for marine black shales from numerous geological periods. They have been related to (1) increased OC flux to the seafloor induced by productivity pulses, (2) climate-induced variations in organic productivity, coupled with variations in bottom water redox [Saelen et al., 2000, and references therein], (3) water column stratification [Tyson et al., 1979], and (4) climate controlled production of expandable clay minerals (smectite-type) into oxygen depleted continental margin settings, catalyzing OC burial via organo-mineral interactions [Kennedy and Wagner, 2011]. Following the conceptual model for the Cretaceous, enhanced runoff would have promoted the establishment of anoxia/ euxinia and black shale formation along the Boreal Seaway through enhanced marine productivity. Conversely, reduced runoff would have resulted in reduced nutrient flux and marine productivity and thus partial re-establishment of oxic conditions in the water column. As demonstrated by the climate model such a situation would correspond to a more northerly position of the Intertropical Convergence Zone (ITCZ) during the late Jurassic. There is a remarkable consistency between the evidence deduced from the climate modeling and the sedimentological record across the British sector. This leads us to suggest that OC burial and black shale deposition in the KCF were intimately linked to precipitation changes that can be associated with orbital time scale fluctuations in the atmospheric circulation, and in particular changes in the extent of the Hadley cells.
Our alternative interpretation of prevailing climatic conditions in the southern part of the Late Jurassic Boreal Seaway implies a maximum reach of the ITCZ as far north as~35°N. Such a scenario recognizes the substantially wider latitudinal shift of the ITCZ in the modern Pacific region compared to the Atlantic region, with the ITCZ migrating seasonally over~60°of latitude and a maximum northern position at~35°N during boreal summer. Identification of the outer subtropical boundary hints toward a low-latitude influence on marine sedimentation and OC burial in the Arctic sector of the Boreal Seaway. Mutterlose et al. [2003] reported OC-rich Kimmeridgian-Tithonian aged shales as far north as the Barents Sea (54°N in the Late Jurassic), with sedimentary cycles containing a distinct precessional-pacing, consistent with some influence from low-latitude fluctuations in insolation. In the late Jurassic the northern Hadley cell may therefore have extended, at least temporarily, imposing dynamic, subtropical conditions close to the paleo-Arctic.
We therefore propose that orbital forcing in the Late Jurassic would have controlled alternations between extremely humid conditions, supporting peak OC burial in the KCF, and dryer periods where freshwater supply and redox conditions relaxed, resulting in lower OC burial. A depositional system with strongly alternating hydrological conditions is consistent with many black shales from the Cretaceous Oceanic Anoxic Events [Wagner et al., 2013;Beckmann et al., 2005] and the Neogene Mediterranean sapropels [Rossignol-Strick, 1985].
The onset of a drier climate in the upper Tithonian of the British sector may indicate a gradual transition to a more subtropical climate with stronger trade winds. The underlying mechanism for this large-scale shift in climate conditions remains unclear, but a link to the long-term and global scale drawdown of CO 2 during peak Kimmeridgian black shale deposition has been proposed [see also Wignall and Ruffell, 1990]. These conditions would have promoted global cooling and eventually incipient glaciation [Donnadieu et al., 2011]. The emergence of incipient south polar ice sheets in the later Tithonian should has led to an increase in global surface meridional temperature gradient and a contraction of the outer boundaries of the Hadley cells, gradually shifting the British KCF out of the direct influence of tropical and eventually also subtropical climate conditions.

Conclusions
The Kimmerdige black shales from the late Jurassic epeiric Boreal Seaway were deposited at high average sedimentation rates (8 cm/kyr) enabling the investigation of paleoclimate and paleoceanography at high temporal resolution and associated hydrocarbon source-rock formation. The KCF was deposited between Paleoceanography 10.1002/2015PA002911 ~35°N and~65°N in temperature-or salinity-stratified subbasins. Deposition was strongly influenced by frequent storm activity and high rates of sedimentation leading to expanded shale sections and massive burial of organic matter. Analyses of the relationships between clay mineral assemblages (kaolinite/illite) and OC content support the conclusion that changes in OC burial resulted from orbitally paced fluctuation in rainfall intensity beneath the ascending limb of the Hadley cell, i.e., under the direct influence of the tropical ITCZ. Our new climate simulations support this tropical scenario. Together, these suggest that the variable but overall very high burial of OC in the KCF was controlled by very similar mechanisms as some of the Cretaceous OAE's and the Mediterranean sapropels.