GENIEfy: creating a Grid ENabled Integrated Earth system modelling framework for the community.

精灵项目的目标是建立简化的、运行更快的地球气候系统模型，使它们更容易使用，并更广泛地供其他想要和需要使用它们的人使用。最先进的气候模型(如优秀的哈德利中心模型)在空间和时间上都工作在非常详细的尺度上，因此又大又慢，除了超级计算机之外，不能用来模拟超过几个世纪的气候，即使这样也可能需要几个月的时间才能给出结果。我们的目标是模拟数千年的气候，因为地球气候在过去经历了数千年(甚至数百万年)的重大和戏剧性的气候变化(包括冰河时代)，我们需要了解这些变化，这样我们才能更有信心地模拟未来的气候。我们的目标是使用更广泛使用的计算机(包括高端PC)来实现这一点，以便让更多不是计算机专家的科学家能够探索他们的想法。为了建立这样的模型，运行速度要快数千倍，我们必须使用更粗糙的网格，提供的结果不那么详细，还必须使用相互作用来控制气候的物理、化学和生物过程的简化版本。这样做的代价是，我们预期的结果不那么准确，但我们可以进行大量的运行来比较结果，从而找出我们可以容忍多大程度的简化，以及答案的准确性，这在大型模型中几乎是不可能的。在精灵的第一阶段，我们已经成功地建立了一个这样的简化模型的小家族，并使用先进的计算方法来调整它们与真实世界的数据，并解决关于它们对我们不太了解的东西的依赖的难题。我们现在的目标是在此基础上再接再厉；(1)试图找出需要多少细节才能达到不同的精度水平；(2)在我们的气候模型与技术发展和经济模型之间建立联系；(3)使我们的模型更类似于大型主流模型(从而更容易与它们的结果进行比较)；(4)使它们更好、更快；以及(5)使它们更容易--被创建者以外的人广泛使用。为此，我们将利用先进的软件和计算技术，使强大的计算机资源可以通过互联网获得。

The GENIE project's aim is to build simplified and faster-running models of the Earth's climate system, and make them easier to use and more widely available to other people who want & need to use them. State-of-the-art climate models (such as the excellent Hadley Centre model) work on quite detailed scales in both space and time, and are consequently big and slow, and cannot be used to simulate more than a few centuries of climate on anything other than a super-computer, and even so may take months to give results. We aim to model the climate for many thousands of years, because the climate of the Earth has undergone major and dramatic climate changes (including ice ages) in the past, over periods of many thousands (and even millions) of years, and we need to understand these so that we can model future climate with more confidence. And we aim to do this using more widely available computers (including top-end PCs) so that more scientists who are not computer experts can explore their ideas. To build such models which run thousands of times faster, we have to work with coarser grids, giving us less detailed results, and also use simplified versions of the physical, chemical and biological processes which interact to control the climate. The price of this is that we expect less accurate results, but we can afford to do large numbers of runs to compare the results, and so find out how much simplification we can tolerate, and how accurate the answers are, which is almost impossible with the big models. In the first phase of GENIE we have successfully built a small family of such simplified models and used advanced computational methods to tune' them to real world data, and to attack difficult questions about their reliance on things that we do not know very well. We now aim to build on this; (1) to try to find out how much detail is really needed for various levels of accuracy; (2) to create links between our models of the climate and models of technological development and the economy; (3) to make our models more similar to the big mainstream models (and so make it easier to compare results with them); (4) to make them better and even faster; and (5) to make them much more easily & widely usable by people other than their creators. We shall do this by exploiting the advanced software and computing technology which makes powerful computer resources available over the internet.