JOURNAL OF SEDIMENTARY RESEARCH,VOL. 71, NO. 1, JANUARY, 2001, P. 15–26 Copyright q 2001, SEPM (Society for Sedimentary Geology) 1073-130X/01/071-15/$03.00 CLOSED-SYSTEM BURIAL DIAGENESIS IN RESERVOIR SANDSTONES: EXAMPLES FROM THE GARN FORMATION AT HALTENBANKEN AREA, OFFSHORE MID-NORWAY FAWAD A. CHUHAN, KNUT BJØRLYKKE, AND CAROLINE J. LOWREY Department of Geology, University of Oslo, P.O. Box 1047 Blindern N-0316, Oslo, Norway ABSTRACT: The Middle Jurassic Garn Formation of the Haltenbanken area has been studied using mineralogical and geochemical data from 21 wells, ranging in burial depths from 2.0 to 4.1 km relative to seafloor (RSF). K-feldspar and plagioclase contents show variations on a re- gional scale both laterally and as a function of burial depth. The con- tent of pore–filling authigenic illite increases sharply, and the content of K-feldspar and kaolinite decreases in Garn sandstones presently at depths greater than 3.6–3.7 km RSF (120–1308C). The depletion in K- feldspar below 3.7 km RSF is not accompanied by lower potassium values in the bulk chemical composition (wt % K 2 O). This suggests that the potassium released during K-feldspar dissolution is retained in the sandstones and is precipitated as illite. The variations in bulk contents of potassium and sodium are therefore considered to be re- lated principally to primary variations in sandstone mineralogy. The shallower sandstones (, 3.7 km RSF) with average wt % K 2 O greater than 0.95 (K/Al molar ratio . 1/3) have a K-feldspar:kaolinite ratio greater than one. The deeply buried (. 3.7 km RSF) sandstones with similar potassium contents contain excess K-feldspar and most of the kaolinite is illitized. However, deeply buried sandstones containing an average of 0.38 wt % K 2 O (K/Al molar ratio , 1/4) contain a significant amount of kaolinite but negligible K-feldspar. This suggests that the K-feldspar:kaolinite ratio before the onset of illitization was less than one, and hence that the kaolinite–illite reaction has been re- stricted by an insufficient supply of potassium (absence of K-feldspar). This illustrates how illitization of kaolinite depends upon K-feldspar as a local source of potassium. Prediction of illitization in sandstones, therefore, must be based on integration of models for provenance, fa- cies, and early diagenesis in addition to burial and thermal history. The formation of pore–filling authigenic illite in these sandstones is an important influence on the total reservoir quality. INTRODUCTION The Middle Jurassic Garn Formation is a significant hydrocarbon res- ervoir sandstone in the Haltenbanken area (Fig. 1). It ranges in thickness from 14 to 114 m, and is thickest in the southwestern area and thins towards the northwest and northeast, reflecting both varying depositional thickness and local erosion during Late Jurassic to Early Cretaceous rift tectonics and resultant block faulting (Ehrenberg 1990; 1992). The structural and tectonic evolution of the Haltenbanken area has been described by a num- ber of authors (Bukovics and Ziegler 1985; Ehrenberg et al. 1992; Blystad et al. 1995). The Garn Formation is the uppermost unit of the three formations that constitute the Fangst Group, which is contemporaneous with the Brent Group of the northern North Sea (Ehrenberg 1990). The depositional en- vironment of the Garn Formation is the subject of some discussion, with both fluvial and marine environments proposed by different authors (Gjel- berg et al. 1987; Harris 1989; Provan 1992). It covers a wide area and forms blanket-like sand deposits that led Gjelberg et al. (1987) to propose a model termed as ‘‘back stepping of progradational cycles’’ in which the Garn Formation forms a series of ‘‘sheet-sand-cyclothems’’ comprising alternating regressive sand wedges and thinner sheets of reworked trans- gressive sand. Overlying the Garn Formation are marine shales of the Mel- ke and Spekk formations, contemporaneous with the Heather and Draupne shales, respectively, of the northern North Sea. The diagenesis of the Garn Formation in the Haltenbanken area has been discussed in a number of publications (e.g., Bjørlykke et al. 1986, Bjørlykke et al. 1989; Ehrenberg and Nadeau 1989; Ehrenberg 1990, 1991). The problem discussed in this paper is related to the factors influencing illitization in the sandstones of the Garn Formation. Illite is an important diagenetic mineral because its fibrous and pore-bridging morphology strongly influences reservoir quality, particularly permeability and oil sat- uration. There is significant disagreement, however, among the various au- thors regarding the causes, timing, and origin of authigenic illite, and in broad terms there are two major schools of thought. One group of authors consider authigenic illite to originate in an open system because of the influx of (hot) potassium-rich fluids into the sand- stones (e.g. Hurst and Irwin 1982; Jourdan et al. 1987). These fluids may be sourced through compactional processes (Glassman 1992; Berger et al. 1997) and may be introduced to the reservoirs along faults (Burley and MacQuaker 1992). In all these cases illite precipitation is caused by the introduction of K 1 from sources external to the sandstone undergoing il- litization. Hence, the sandstone system is classified as open by these au- thors. Sandstones usually show declining K-feldspar content with depth both in the North Sea and the Gulf Coast basins (e.g., Ehrenberg and Nadeau 1989; Bjørlykke et al. 1992). Some authors have suggested that dissolution of K-feldspar is accompanied by loss of K 1 from the sandstones and that potassium can be transported for relatively long distances through a sedimentary sequence (Harris 1992; Gluys and Coleman 1992; Milliken et al. 1994; Sutton and Land 1996). It has also been suggested that shales act as sinks for the K 1 released from the K-feldspar in sandstones (Ohr et al. 1991; Awwiller 1993; Furlan et al. 1996). The second group of authors consider the formation of authigenic illite to occur within a relatively closed system at 3.5–4.0 km (120–1408C) at the expense of kaolinite and utilising K 1 from K-feldspar (Bjørlykke 1983; Ehrenberg and Nadeau 1989; Bjørlykke and Aagaard 1992; Aagaard et al. 1992; Bjørlykke 1998). There is clearly a lack of consensus among authors about the degree to which potassium and other ions can be transported during burial diagenesis. A relatively isochemical model for burial diagenesis implies that the com- position and reservoir quality of sandstones are functions of the initial sediment composition and early diagenesis. Reservoir quality is then linked to provenance and facies in addition to thermal history. In a more open system the reservoir quality is less dependent on provenance and facies, and more on longer-distance diffusive and advective transport. The purpose of this paper is to study the relationship between the mineralogical and geochemical composition of the Garn sandstones, their burial depths, and the degree of illitization. We also investigate whether varying potassium contents in the deeper wells is a diagenetic feature that reflects potassium loss or a result of variations in the depositional mineralogy of the sand- stones. DATABASE AND METHODS This study is based on petrologic modal analysis (300 points/thin section of 319 samples), whole–rock chemical analysis (308 samples), XRD bulk– rock analysis (329 samples), and XRD analysis of clay separates (329 sam- ples) from 21 wells (Tables 1–3 at URL http://www.ngdc.noaa.noaa.gov/ mgg/sepm/jsr/). The database utilised includes analyses performed by the authors as well as data and samples provided by other authors (Bjørlykke et al. 1986; Ehrenberg and Nadeau 1989; Ehrenberg 1990, 1991; Olsen 1996). Selected samples were also studied using the scanning electron mi-