Comparison of CMlP6 and CMIP5 simulations of precipitation inChina and the East Asian summer monsoon
Projections of precipitation over China based on CMIP6 models
Insight from CMIP6 SSP-RCP scenarios forfuture drought characteristics in China
Does CMIP6 Inspire More Confidence in SimulatingClimate Extremes over China?
Evaluation and Projection of Surface Wind Speed Over ChinaBased on CMIP6 GCMs
Projected drought conditions in NorthwestChina with CMIP6 models under combinedSSPs and RCPs for 2015-2099
Online learning assistance support, [contact the editor, fill in the feedback form in detail] --- very important
Topic 1 Model comparison plan in CMIP6
1.1 Introduction to GCM The Global Climate Model (GCM) , also known as the Global Circulation Model or the Global Atmospheric Model, is a numerical model used to simulate the Earth's climate system. Such models use a series of mathematical formulas to describe the major components of the climate system, including the atmosphere, oceans, frozen soils, and biogeographic processes on land and ocean surfaces.
The accuracy of GCM in space and time can be adjusted according to the needs, usually the resolution can be from hundreds of kilometers to several kilometers, and the time step can be from minutes to hours.
1.2 Introduction to CMIP
CMIP, short for the Climate Model Intercomparison Project (Climate Model Intercomparison Project), is an international cooperation project initiated by the World Climate Research Program (WCRP). Its purpose is to understand past, present and future climate change by collecting and comparing simulation results from various global climate models (GCMs).
1.3 Introduction of Related Comparison Plans
Topic 2 Data Download
2.1 Method 1: Manually
Use the official website
2.2 Method 2: Automatically use Python command line tools
2.3 Method 3: Semi-automatic shopping cart
Use the official website
2.4 Cropping netCDF file
Crop netCDF format based on QGIS and CDO
Operations in QGIS
cropping effect
2.5 GCM whose processing date is not 365 days takes BCC as an example
Topic 3 Basic knowledge
3.1 Basics of Python
Python is a high-level, interpreted programming language with a clean and concise syntax for rapid development. In atmospheric science, Python is favored for its rich scientific computing and data analysis libraries. These libraries, such as Numpy, Scipy, Pandas, and Xarray, provide powerful support for processing atmospheric science data.
●Numpy: Numpy is the core library for scientific computing in Python, providing high-performance multidimensional array objects and related tools. For the processing of atmospheric science data, such as temperature, pressure, wind speed, etc., multidimensional arrays are usually used. Numpy provides a rich library of functions to process these arrays, including mathematical operations, logical operations, shape operations, sorting, selection, and more.
●Scipy: Scipy is a Python-based open source software for numerical integration and numerical solution of differential equations in scientific computing, linear algebra, optimization, signal processing, etc. In atmospheric science, such as performing Fourier analysis on data such as air temperature and air pressure, and solving partial differential equations in atmospheric dynamics, etc., can all be implemented using Scipy.
●Pandas: Pandas is built on Numpy, which makes data cleaning and analysis work faster and easier. Pandas is specially designed for handling tabular and mixed data, while Numpy is more suitable for handling uniform numerical array data. Pandas is a very useful tool in atmospheric science, such as time-series analysis of observation data from weather stations, handling mixed types of meteorological data, and data cleaning, filtering, and statistical operations.
3.2 Basic operation of CDO
CDO (Climate Data Operator) is a climate and meteorological data processing tool commonly used in the field of atmospheric science. It is a powerful command-line tool that can process and analyze gridded and gridless data, and supports a variety of data formats, including netCDF, GRIB, SERVICE, EXTRA and IEG. CDO provides a rich set of function libraries that can be used for various common data operations,
Including: ●Basic operations: such as selecting, extracting and modifying variables, dimensions, attributes, etc.
● Numerical operations: such as four arithmetic operations, statistical operations, function operations, etc. For example, the mean, maximum, minimum, standard deviation, etc. of the data can be calculated. ● Spatial operations: such as re-grid, interpolation, summary, selection and extraction of geographic areas.
●Time operation: such as selecting and extracting time period, calculating time average or accumulation, etc.
3.3 Basic operation of Xarray
Xarray is a Python library for processing multidimensional array data. It provides a series of high-level interfaces for data manipulation and analysis based on numpy, and can well support netCDF, a network-based self-describing data format. Therefore, it is widely used in atmospheric science and climate science.
Key features of Xarray include:
●Label-based data manipulation: Xarray uses dimension names instead of axis numbers for data selection and manipulation, which greatly enhances the readability and maintainability of the code.
●Automatic alignment of data: Xarray can automatically align variables and coordinates of different data sets when performing operations. ●Group operation and data pivot: Xarray supports group-by and pivot functions similar to pandas.
● l/O operation: Xarray provides very good support for various data formats, especially for reading and writing netCDF data.
Topic 4 Single-point downscaling
4.1 Delta method
Delta method (Delta Change Method), also known as incremental method or difference method, is a simple and commonly used method for climate model downscaling. This approach assumes that the magnitude of climate change will remain constant in the future relative to historical periods. Therefore, for a specific future period, the future climate state can be estimated by calculating the difference between the past and present climate (ie delta) and applying it to future climate predictions. This method can be applied to the prediction of climate variables such as temperature and precipitation.
4.2 Statistical corrections
A revision of the Probability Density Function (PDF).
The basic idea of this method is to obtain more accurate small-scale climate variables by modifying the PDF output by the large-scale model to make it more consistent with the PDF of the observation data.
4.3 Machine Learning Methods
Downscaling is the process of converting coarse-scale global climate model (GCM) output data to a finer scale at the surface. Machine learning methods have been successfully applied to downscaling techniques due to their powerful capabilities in dealing with complex pattern recognition and high-dimensional data problems. In the field of climatology, machine learning has been successfully used to correlate coarse-scale climate model outputs (e.g., temperature and precipitation) with other environmental variables (e.g., topography and soil type) to obtain higher-resolution climate predictions.
Implementation steps
● Build features
● Modeling
●Model evaluation
4.4 Multi-algorithm integration method Multi-algorithm integration
Bayesian Model Averaging (BMA)
Bayesian model averaging is a statistical method for determining the posterior probabilities of various models based on observed data. As opposed to choosing a single best model, Bayesian models consider all possible models on average and then take a weighted average based on the posterior probability of each model. Python+pymc3 implementation
Topic 5 Regional Downscaling of Statistical Methods
5.1 Delta method
5.2 Based on the probability correction method
Topic 6 Dynamic downscaling based on WRF model
Dynamic downscaling typically uses higher-resolution regional climate models (RCMs) that are driven by larger-scale global climate models. Among them, the WRF (Weather Research and Forecasting) model is currently one of the most widely used regional climate models. The WRF model is a flexible, general circulation model suitable for climate and meteorological studies at various scales. Its main feature is that it has high resolution (up to several kilometers), and it can take into account many important geophysical processes, such as cloud formation, precipitation, land surface processes, ocean processes, boundary layer processes, radiation, chemical processes, etc. .
6.1 Preparation of WRF driving data for CMIP6
Use the cdo tool to recode the gcm output file to prepare the wrf drive data
6.1.1 Data preparation for the pressure coordinate system
6.1.2 GCM data preparation for sigma coordinate system
6.1.3 WPS processing
6.2 WRF mode operation
6.3 Post-processing of the model
● extract variables
●Statistics of variables
●Visualization of variables
Topic 7 Typical Application Cases - Climate Change 1
7.1 Downscaling for wind speed
7.2 Downscaling for shortwave radiation
Topic 8 Typical Application Cases - Calculation of Climate Change 2ECA Extreme Climate Index
ECA (European Climate Assessment) is a European climate assessment project that publishes a series of extreme climate event indices around the world. These indices are widely used in climate change research, especially in studying extreme weather and climate events.
The extreme climate indices of ECA mainly include the following categories:
Temperature indices: These indices are mainly used to measure temperature extremes, such as the number of hot days (TX90p, the number of days with the highest temperature in the year above the 90th percentile), the number of cold days (TN10p, the number of days with the lowest temperature in the year below the 10th percentile The number of days), the number of hot nights (TN90p, the number of days with the lowest temperature in the year exceeding the 90th percentile), the number of cold nights (TN10p, the number of days in the year with the lowest temperature below the 10th percentile), etc.
Precipitation index: These indices are mainly used to measure the extreme conditions of precipitation, such as the maximum continuous 5-day precipitation (RX5day), the number of precipitation days greater than or equal to 10mm (R10mm), the number of precipitation days greater than or equal to 20mm (R20mm), precipitation intensity ( SDII), etc. These indices are important for understanding and predicting the impact of extreme climate events, as extreme weather events (such as heat waves, droughts, floods, etc.) tend to have a greater impact than average climate change. Therefore, research on these indices can help us better understand and adapt to climate change.
lConsecutive dry days index
lConsecutive frost days index per time period
lConsecutive summer days index per time period
lConsecutive wet days index per time period
Topic 9 Typical Application Cases - Estimated Start and End Time of the Growing Season in the Ecological Field
1. Establish the start and end of the growing season in meteorological data and VIPPHEN remote sensing phenology data
2. Estimate the start, end and length of the long growing season under future climate scenarios
Topic 10 Typical Application Cases - Hydrological and Ecological Model Data
● SWAT data preparation
●Biome-BGC data
Biome-BGC is a model that uses site description data, meteorological data, and vegetation physiological and ecological parameters to simulate daily-scale carbon, water, and nitrogen fluxes. The spatial scale of its research can be extended from point scales to terrestrial ecosystems. In the case, the meteorological data of CMIP6 are prepared by single-point simulation.