Rainfall forecasting is a very important problem in the field of hydrology and meteorology. Especially, short-term rainfall forecasting is closely related to resident's daily life. For example, forecasting the situation of stagnant water on the road, providing weather guidance for the flight, and providing short-term heavy rainfall warning in the city.However, existing solutions achieve low prediction accuracy for short-term rainfall forecasting. Numerical forecasting models can achieve overall accuracy but always perform worth in some short-term conditions. Data-driven approaches always neglect the influences of physical factors in upstream or downstream regions, which lead to the forecasting accuracy fluctuates in different areas. Rainfall forecasting is affected by many factors, such as high-altitude physical factors and surface factors.High-altitude physical factors play important roles in the movement of rainfall system. Surface factors on the Earth also cause different rainfall. Difference surface factors represent different factors between region and surrounding area. Therefore, it is very reasonable to forecast rainfall by studying the relations between high-altitude physical factors, surface factors and rainfall. In this project, a Dynamic Regional Combined short-term rainfall forecasting model (DRCF) using Multi-Layer Perceptron (MLP) is proposed. The input of the model includes five high-altitude factors and seven different surface factors. In summary, we have addressed a series of techniques challenges in this work, and the central contributions are summarized as follows: 1.Principle Component Analysis (PCA) is used to determined the input of of MPL and a special greedy algorithm is proposed to determine the suitable structure of MLP. 2.The relation between forecasting area and surrounding area is established using MPL. Based on the relation, a dynamic regional combined rainfall forecasting model is proposed. The strategy of dynamically adjusting area is also enhanced to effectively improve prediction accuracy. 3.The model is finally validated by a large number of meteorological site data, including 56 sites, and these sites are distributed in all parts of China.

flow_modeling

A python algorithm that models surface water flow, given either random or real-world elevation and rainfall datasets.

forestclaw

Quadtree/octree adaptive PDE solver based based on p4est.

fupral

Fortran OO implementation of a generic container using an unlimited polymorphic class. Implementation of a resizable container array and a double linked list.

fvcom

A prognostic, unstructured-grid, finite-volume, free-surface, 3-D primitive equation coastal ocean circulation model developed by UMASSD-WHOI joint efforts

getsomecats

gpt_academic

为ChatGPT/GLM提供图形交互界面，特别优化论文阅读润色体验，模块化设计支持自定义快捷按钮&函数插件，支持代码块表格显示，Tex公式双显示，支持Python和C++等项目剖析&自译解功能，PDF/LaTex论文翻译&总结功能，支持并行问询多种LLM模型，支持清华chatglm等本地模型。兼容复旦MOSS, llama, rwkv, 盘古, newbing, claude等

hydraulic-and-hydrologic-modelling-tool

HHMT is an integrated package of hydraulic and hydrologic analysis tools, in which the user interacts with the system through the use of a graphical user interface (GUI). The system is capable of performing steady and unsteady flow water surface profile calculation and several hydraulic computation

gfjykldd Goto Github PK

Zhou Qi's Projects

Recommend Projects

Recommend Topics

Recommend Org