Abstract

Seismic models enhanced with well logs help to locate sand concentrations. Many well log curves contain patterns with very high frequencies compared to seismic responses. Quantitative analysis of log patterns can create large numbers of sand vectors for plots that resemble and complement 3D seismic attribute maps. Complex log patterns and improved predictions about changes in log responses over distances have been integrated with broad geophysical coverage. Detailed and extensive reservoir models were created.

Introduction

Finding and modeling sandstone reservoirs can be difficult on any continent. However, Texas, Eastern New Mexico, and other mountain states extending to the north-northwest have areas with especially difficult to decipher variations in sandstone deposition. Even a single county may have distinctly different fields with geology that includes marine sediments, estuaries, several different types of deltas, and fluvial features. Other complications include major changes in sediment flow directions that occurred within a relatively small amount of geologic time. In addition, depths of burial can limit seismic detection of interbedded reservoirs.

Furthermore, general shortages of well logs often mandate that as much useful geologic information as possible should be extracted from each log. Techniques were needed to quantitatively utilize sand patterns on logs to improve sand maps.

Theory

High-energy pulses associated with water movements often radiate dissipating forces capable of influencing fine-grained sedimentation patterns that often extend for several miles. Consequently, log wavelets in shales may retain signatures relatable to sandstone reservoirs in the area. For many years, faint, almost shaled out sand responses on logs have been observed and used for correlations and indicators of better developed sands at distant locations. Qualitative analysis of log signatures has been done for decades. Shapes on spontaneous potential (SP) curves may be related to bars, braided streams, channels, and other deposits.

Quantitative analysis of log patterns has the capability to provide better reservoir geometry predictions. Directional variations in sediment transport equations can be determined from rates of change in log responses between wells. In other words, directions and magnitudes (vectors) can be extracted from common lithology log responses. Large numbers of vector solution points calculated from each well log can be plotted to model sediments within the surrounding area. In contrast, the conventional technique of averaging many log responses into a single sand thickness largely wastes huge amounts of detailed data. Standard sand mapping techniques are not being discouraged. Instead, log derived sediment vector plots can be created to compliment sand maps and reveal geologic details. Variations in seismic trace patterns, extracted pseudo-sonic logs, and depositional models also provide transport information.

Generally consistent mathematics can be used to describe specific components of sediment transport mechanisms. Linear equations have been frequently derived for log response changes aligned with the flows of streams and channels. Logarithmic variations are common for sediments altered by long shore currents and some wave conditions. There are many fine statistical geology publications that can provide general geometries when only a few logs are available. However, when possible, one should develop local knowledge based on adequate quantities of actual log measurement changes.

This content is only available via PDF.
You can access this article if you purchase or spend a download.