Field development planning based on integrated studies, including construction of a 3D geo-model, is an accepted standard for development of larger assets. The need to keep field development plans ‘evergreen’ is well recognized, but often difficult to realize in practice. This paper will focus on the results of a drilling campaign that followed an integrated study previously described in SPE paper 62901(Herweijer et al., 2000). As with many drilling campaigns, the results covered a range from very positive to negative outcomes. This paper will show how the result of the campaign, which was an overall success, was affected by:The modeling process, which assumed a base case with perturbations. It became anchored on a specific, optimistic, fluid contact scenario, rather than including a realistically wide range.The process of multi-disciplinary work, which, although rigorous, did not in itself guarantee that a realistic range of scenarios had been covered.Changes in team membership prior to execution of the infill campaign. This lead to a partial lack of ownership, resulting in the missing of signposts that could have lead to re-tuning the campaign by accounting for more downside scenarios.The relatively elaborate study and modeling process, which discouraged updating and hence lead to poor integration of some early negative drilling results into the model. The conclusion of the paper is that the whole study became somewhat ‘model-centric’. The model served as an unbeatable ‘truth’: trusted more than pre-model mapping, not fully connected to post-model ‘mass production’ infill drilling results, and used on problems to which it was not suited. As a result of the lessons learned we have been able to assess the real value of knowledge continuity and model ‘ever-greening’ from early stages of development through to later ‘harvest’ drilling. Introduction With the upsurge of 3D modeling as de-facto standard for integrated subsurface studies, there is much scope to address managerial issues that directly impact value delivered by development following such studies. We present a case regarding management of such a study and lessons learned for future modeling-driven development activity, and present how improvement of study management can directly impact value. This paper is a follow-up of paper SPE 62901 (Herweijer et al., 2000), which presented the results of a 3D model study of the onshore Moomba North gas field in Australia. This study was part of a larger effort of several integrated studies that were conducted in parallel at a relatively fast pace to underpin further development of a suite of gas field assets in the Cooper Basin in South Australia (Bard et al., 2000). We note that in the current high activity setting of the oil and gas industry, this fast pace parallel study approach is often practiced. The original field development plan (FDP) for Moomba North, called for 20–30 development wells, some of them clearly crestal infill wells, with others extending a fairly dense well pattern towards the flank of the field (Herweijer et al., 2000). During 2000 and 2001, a total of 32 wells were drilled in two campaigns. From the start it was recognized that drilling outcomes would be variable, and that in the end the total campaign results would be the economic benchmark.

Discontinuous fluvial sands with highly variable permeability patterns pose a significant challenge for further development of mature onshore gas fields in South Australia. To obtain an optimal strategy for further infill drilling a multi-disciplinary reservoir characterization study was undertaken. This study combined all available data (petrophysics, geology, geophysics, and engineering) into a 3D stochastic geo-model before various upscaling methods could be considered for creating a reservoir simulation model. Coarse upscaling methods were used to match macro-scale reservoir characteristics (e.g. the effective permeability in the well's drainage area) while smaller scale heterogeneity (RFT pressure measurements) were matched using permeability predictors that preserved the observed variability from core data. Several geological scenarios were designed for the model. These scenarios included various assumptions on sand content of the main producing horizon, sandbody dimensions, the permeability distribution, and continuity of coals acting as vertical baffles. During history matching, some of these scenarios were tested as alternatives in pursuit of a history match. Scenarios were refined/rejected and/or iterations made to re-assess the data. The history matching procedure revealed the significance of outlier permeability streaks as conduits to effectively drain reserves. The study was conducted under the constraints of a fairly tight operational schedule to meet a development drilling deadline. Drilling locations and the associated economic evaluations were based on simulated forecasts from a number of stochastic realizations/scenarios while connected volume calculations were made for a larger number of realizations. Introduction We present a case study of a multi-layered gas field in the South Australian Cooper basin. The field consists of discontinuous fluvial-deltaic sands. Thirty development wells were drilled after the initial discovery, at a well spacing of approximately 500–1000 acres. Drilling locations were guided by structural trends and geological insights regarding sand distribution obtained from appraisal wells and early development wells. Accurate replication of reservoir performance and characteristics required the resolution of two primary issues; that ofconnectivity of the various fluvial sand facies andthe permeability distribution, and the geological (facies, diagenesis) control on that permeability distribution. These factors contribute to the observed non-uniform pressure distribution, areally and vertically. The scheduled completion date of the study had to be strictly adhered to due to operational requirements. The technical and economic success of the expected infill campaign relied upon wells being able to penetrate thick sand prone sections, yet these sands have to be sufficiently disconnected from neighboring wells to ensure they are not already depleted. In this paper we describe the workflow, applied to construct the 3D models, and the results of the simulation study. The paper addresses various geological and petrophysical scenarios used for risk management and to capture the uncertainty within the data and the interpretation. At the time of writing this paper several new wells of the infill campaign have been drilled. The last section of this paper compares actual reservoir parameters to the predictions of the 3D models. Geological setting The modeled section consists of the Permian Toolachee and Daralingie Formations deposited in a continental fluvial-deltaic-lacustrine setting.