Creating a learning community through knowledge management: the Mandala Project

The work of a forecaster is complex, categorised by uncertainty, incomplete information, multiple sources and a great variety of data and strict timeless all overlaid by a legal regime. Forecasters are required to exercise judgement because science is often inadequate at the level of detail required by specific forecasts. In such an environment, work activity assumes not only task performance, construction of a forecast, but also the review and re-assessment of the work done in order to understand and learn from the experience. The exercise of judgement and review together represent ‘knowledge work’, which requires an enhanced decision support approach which we term Intelligent Decision Support (IDS). In this paper we present a knowledge management framework in the context of intelligent decision support for weather forecasting. We illustrate the implementation of our knowledge management framework with a system for collaborative weather forecasting. The system provides an environment in which the forecaster is able to explore all available data, multiple it, access guidance material, construct a forecast and provide a rationale for the forecast. The system intelligently manages the data allowing the forecaster to exercise their judgement. The collaborative aspect involves forecasters being able to reconcile their individual forecasts with other forecasters’ work. This is achieved by making the forecast and its rationale, including the guidance used in the forecast, available to other forecasters. With these facilities we are also able to demonstrate how individual and organisational learning processes are supported. The Mandala project presents a unique environment to apply and extend our knowledge management approach. The Mandala Project represents the third wave of development of integrated support facilities for operational forecasters. It aims to provide online services that include access to wide variety of data, data visualisation and graphical editors, historical forecast databases and automated forecast guidance amongst other services.<div><br></div><div>Presented at: IFIP WG 8.3 conference Decision support through knowledge management; 2000 Jul 9-11; Stockholm, Sweeden. p. 190-208