Abstract: The mid-to-deep reservoirs of the Miocene Huangliu Formation in the F gas field of Yinggehai Basin are dominated by complex gravity flow deposits. The primary producing interval, namely the Unit-a of submember Ⅰ of Member 1 of Huangliu Formation, features small thickness and pronounced heterogeneity in physical properties. To address the ambiguous geological implications of seismic amplitude attributes and challenges in sweet-spot reservoir prediction, combined spectral expansion was employed to enhance the resolution of original seismic data. This was integrated with fine-scale thickness characterization via broadband inversion, and three thickness gradients were partitioned using λ/8 and λ/4 as boundaries. Subsequently, by utilizing effective high-frequency seismic data volume derived from multi-wavelet decomposition and reconstruction, high-frequency amplitude thresholds for effective reservoirs were established. This methodology enabled the semi-quantitative prediction of physical properties for thin reservoirs with thicknesses ranging between λ/8 and λ/4. The results indicate as follows. (1) The optimized combined spectral expansion technology based on spectral shaping balances the enhancement of thin reservoir resolution with the preservation of original sand-body contact relationships, making it suitable for identifying thin reservoirs in complex gravity flows. (2) The amplitude of thin reservoirs is influenced by the coupling of thickness and physical properties, and effective high- frequency information within seismic signals serves as a significant indicator of thin reservoir physical properties. (3) Reservoir conditions vary significantly across different thickness gradients in gravity flow deposits. High-quality reservoirs can be effectively identified by optimizing frequency bands sensitive to physical properties based on the planar distribution of thin reservoirs. The study concludes that high-frequency-indicated prediction results for thin reservoir physical properties based on thickness gradient partitioning are reliable, having effectively guided the deployment of development wells and providing an important reference for the physical property prediction of thin reservoirs in analogous oil and gas fields.

Key words: thickness gradient, effective high-frequency seismic data volume, thin bed, physical property prediction, complex gravity flow, Yinggehai Basin, Miocene, Huangliu Formation

摘要： 莺歌海盆地中—深层F气田中新统黄流组储层为复杂重力流沉积,其主力产层黄流组1段Ⅰ亚段a气组的储层厚度薄、物性非均质性强。针对地震振幅属性指示意义不明,优势储层难预测的问题,在组合拓频提高原始地震资料分辨率的基础上,结合宽频反演精细描述薄储层的厚度,并以波长(λ)的1/8和1/4为界划分3个厚度梯度,然后利用多子波分解与重构得到的有效高频体数据确定有效储层的高频振幅门槛,实现了对厚度梯度在λ/8和λ/4之间的薄储层物性的半定量预测。研究结果表明:①基于谱整形的优化组合拓频处理技术兼顾薄储层分辨率的提升与原始砂体接触关系的保持,适用于复杂重力流的薄储层识别;②薄储层的振幅受厚度、物性耦合影响,地震信号中的有效高频信息对薄储层物性具有明显的指示作用;③重力流储层的不同厚度梯度所对应的储层条件差异较大,在明确薄储层平面展布的基础上优选物性敏感特征频段,能较好地识别优质储层。研究认为,基于厚度梯度划分的高频指示薄储层物性预测结果可靠,有效指导了开发井位的部署,对同类油气田的薄储层物性预测具有重要参考作用。

关键词: 厚度梯度, 有效高频体, 薄层, 物性预测, 复杂重力流, 莺歌海盆地, 中新统, 黄流组