Geochemical features of and source rock contributions to mixed-source oil reservoirs in the 1st member of the Shahejie Formation， Huanghua Depression， Bohai Bay Basin

The Shahejie Formation in the Huanghua Depression boasts abundant mixed-source hydrocarbons. Determining the charging ratio for mixed-source oil reservoirs is critical to the prediction of hydrocarbon reservoir distribution in the depression. Through the analysis and tests of total organic carbon （TOC） content， rock pyrolysis， and gas chromatography-mass spectrometry （GC-MS） experiments on saturated hydrocarbons， we systematically examine the geochemical characteristics of source rocks and crude oil in the formation. A multi-parameter quantitative model for determining the charging ratio for mixed-source oil reservoirs is established by combining hierarchical cluster analysis （HCA） and principal component analysis （PCA）. With this model， we quantify the contributions of different source rocks to the formation of mixed-source oil reservoirs therein. The results indicate that the source rocks in the 3rd member （Es3； the Sha 3 Member） and the 1st member （Es1； the Sha 1 Member） of the Shahejie Formation show moderate to high organic matter abundance and mixed kerogen types of Ⅱ2 and Ⅱ1， respectively， suggesting the presence of mature source rocks. Oil-source rock correlation reveals that Class A crude oil is derived from source rocks in the Es1 and formed in high-salinity， reducing environments， with organic matter originating primarily from lower-class aquatic organisms. Class B crude oil originates solely from source rocks in the Es3 and is formed in medium-salinity， weakly oxidizing to weakly reducing environments， with organic matter sourced mainly from terrestrial plants. Class C crude oil represents mixed oil derived from source rocks in both Es1 and Es3. Calculations using the quantitative model indicate that the contribution rates of source rocks in the Es1 and Es3 to the mixed-source oil reservoirs range from 22.73 % to 59.11 % and from 40.89 % to 77.27 %， respectively. Model validation using laboratory artificial oil mixing experiments shows prediction errors ranging from 0.44 % to 9.94 % （average： 5.26 %）， indicating high predictive accuracy.