在线和离线轨迹计算之间的主要区别是,在线实现作为模型的一部分,和轨迹模型输出的一部分。离线轨迹计算是一个单独的软件,以欧拉模型输出字段作为输入,计算轨迹。在线粒子跟踪的主要优势是,示踪粒子的位移或可以评估每个时间步,不管它有多小。作为一个结果,你可以获得一个真正的粒子轨迹基于模型产生。在线跟踪的缺点是,用户需要知道粒子释放位置和时间的运行模式。下面列出的模型(至少)支持在线拉格朗日轨迹计算:*(混合协调海洋模式(HYCOM)][1]是一个静水,hybrid-coordinate海洋环流模型,粒子计算内置;*(麻省理工学院大气环流模式(MITgcm)][2]是一个中尺度环流模型与许多物理参数化方案和一个动态内核的大气和海洋;*(区域海洋建模系统(rom)][3]是静水,σ坐标(地形跟踪)海洋模型与许多物理参数选项一个庞大的用户社区。另一方面,离线粒子跟踪只能使用欧拉字段一样经常输出。因此,将会有一个错误在计算轨迹取决于输出的频率。 On the other side, the main advantage of offline calculation is that the user can experiment and play around with deployment locations, times, area of coverage etc., because trajectory calculation typically takes much less computational effort than the actual model. There are several options here: * [Connectivity Modeling System (CMS)][4] takes gridded NetCDF files as input and produces Lagrangian trajectory output. Can handle fields from various Arakawa-staggered grids, and may produce probabilistic trajectories, diffusivity specified by user (optional). Compatible with at least HYCOM and ROMS output. * [Ariane][5] also takes gridded NetCDF files and is compativle with at least HYCOM, ROMS and NEMO models. * [TRACMASS][6] seems to be compatible with many circulation models. It also has a [Python interface][7]. All models and tools listed in this answer are freely available for download under different open source licenses. [1]: http://hycom.org [2]: http://mitgcm.org/ [3]: http://www.myroms.org [4]: https://code.google.com/p/connectivity-modeling-system/ [5]: http://stockage.univ-brest.fr/~grima/Ariane/ [6]: http://tracmass.org/ [7]: https://github.com/kthyng/tracpy