Assignment 1 Answer

Achieve 80% to pass

1. Numerical Weather Prediction models are based on:
   1. Artificial Intelligence (AI)
   2. Statistical machine learning from a training set data
   3. Physics laws and parameterization of unresolved physical processes
   4. None of the above

2. The governing equations of a numerical weather prediction model typically
   1. involves the rate of change, with respect to time, of atmospheric state variables
   2. involves spatial operators of atmospheric state variables like gradient
   3. does not couple different atmospheric variables like potential temperature and velocity
   4. All of the above
   5. a and b only

3. After the governing partial differential equations are discretized spatially and temporally, the numerical scheme
   1. has the evolution of state variables at a grid point depends only on the grid point itself
   2. requires the use of data at neighbouring points when computing the evolution of state at a grid point
   3. uses deep learning techniques to deduce the evolution of atmospheric states
   4. evolves the system backward in time to deduce the initial condition

4. Which statement regarding the unstructured grid used in the MPAS-A global model is incorrect?
   1. It must consist of some heptagons
   2. It must consist of some pentagons or polygons with fewer sides
   3. It supports gradual resolution transition
   4. Positions of cell centers and edges are used in the model

5. Regarding the need for Initial Condition (IC) and Lateral Boundary Condition (LBC) data:
   1. A global model does not need IC data
   2. A global model needs LBC data
   3. A regional model needs both IC and LBC data
   4. A global model uses IC data to drive a regional model along the simulation

6. When water vapour in a thunderstorm condenses in an air column,
   1. The atmospheric pressure near surface becomes high
   2. The atmospheric pressure near surface becomes low
   3. The air will be sinking and diverge near surface
   4. Mass continuity will not be kept

Because condensation releases a lot of latent heat, warmed up air and make it rise, removing air molecules near the surface and creating low pressure.

7. What are possible ways to use fine resolution for an area in a numerical weather prediction model?
   1. Use a regional model with several levels of domain nesting, driven by a global model
   2. Use a variable-resolution global model with unstructured grid and smooth resolution transition
   3. Use a global model with domain nesting embedded
   4. All of the above

8. When you use higher level programming language and packages, generally,
   1. you writes more lines of code for a task
   2. the software package dependencies are simpler
   3. you writes code in a more imperative style 
   4. you make bigger foot-print when running it

9. What is not true about a netCDF file?
   1. It contains a set of variables
   2. Each variable is bound to certain dimensions
   3. You can always convert it to a CSV file and use popular spreadsheet software to check its data
   4. You may use the xarray package in Python to open it.

The number of dimensions may exceed two and cannot be arranged as spreadsheet data.

10. Which of the following statement is true (for the existing version of xarray introduced):
    1. If you do not mention the chunks argument when opening a dataset, xarray automatically do it for you and utilizes multi-core computing
    2. If you do not specify the value of the chunks argument when opening a dataset, xarray automatically determine it for you and utilizes multi-core computing
    3. You better have some idea on which dimension to decompose and its chunk size and specify them in the chunks argument when opening a dataset, to leverage multi-core computing effectively
    4. There is no hint on whether multiple CPU cores are being used when using xarray in Jupyter Notebook.