We want to understand whether cloud seeding has the potential to increase rainfall in Qatar. This project will use radiosonde data to identify periods where the atmosphere contains sufficient clouds that could be targeted for cloud seeding. You will use a cloud parcel model intiialised with aerosol particles and hygroscopic cloud seeding nuclei to assess whether the hygroscopic nuclei has the potential to increase surface rainfall.
The code repository for the Bin Microphysics Model is available at https://github.com/UoM-maul1609/bin-microphysics-model. This model is written in Fortran and has some associated Python scripts for analysis and plotting.
Prof. Paul Connolly will start up the VM, give you your username, SSH key (which is a file) and the IP address of the server (VM). To login you need to open CMD (on windows), terminal if on a mac or Chromebook. You would then type:
ssh -i id_virtual_students.key -X <username>@<IP-address>
The first time you log in you may need to answer 'yes' to a question about connecting.
Once logged in you will be interacting the VM in the same way as you would a linux server through the terminal. To download the bin microphysics model type:
git clone https://github.com/UoM-maul1609/bin-microphysics-model
This should download the code to the folder bin-microphysics-model
You need to be in the bin-microphysics-model folder to compile the code. Type
cd bin-microphysics-model
in the terminal / cmd window to enter the folder.
We need to then tell the computer to build the machine code executable that the computer CPU understands. This can be done by typing:
make NETCDFLIB=-L/usr/lib/x86_64-linux-gnu/ NETCDFMOD=/usr/include/ FFLAGS='-O2 -w -o' FFLAGS2='-O2 -w -o' FFLAGSOMP='-O2 -w -o'
The make command tells the computer to use a file called Makefile which is in the bin-microphysics-model folder you downloaded. In this file are rules for compiling the code. There are some additional variables that we pass into the Makefile. These are NETCDFLIB, NETCDFMOD, FFLAGS, FFLAGS2 and FFLAGSOMP, which overrides the values they are set to in the Makefile. They need to be altered for our environment.
You will need radiosonde data, which gives the thermodynamic conditions of the atmosphere close to Qatar. The script download_data_wyoming.py will automatically download data from the Wyoming database at https://weather.uwyo.edu/upperair/sounding.html. You could download each radiosonde file individually, but that would be very time consuming. The script will download a years worth of data for both 0 and 12 local time. Run the script as follows
python3 download_data_wyoming.py
It will take a while to complete, but will download the data into a folder called SondeData in your home folder. You might want to check this folder is empty before you start.
Once downloaded we can read in all of the radiosonde data using the supplied script readSondeData.py. In order to look at the data interactively you will want to run this script from within ipython3. If you just pass the script to the python3 interpreter it will read the files and the exit, so all information will be lost. So type the following:
ipython3
Once iPython has started you can run the readSondeData.py script as follows:
run -i readSondeData.py
This script will read all of the radiosonde files at ~/SondeData into a python list called dats. You can find how many files there are by typing
len(dats)
It probably wont be 2 x number of days in a year because there will be some missing files, which happens when there is a problem with the radiosonde on a particular day. Do not worry about this.
We generally extract one of these radiosonde ascent files to look at in more detail. Before we do that lets look at some general metrics for the data we have just read in.
When we read in the radiosondes using readSondeData.py it puts the following variables into memory within iPython (you can type who or whos to see them)
time
tlcl
plcl
cape
cin
pw
These are: (1) time; (2) temperature of the lifting condensation level; (3) pressure of the lifting condensation level; (4) Convective Available Potential Energy; (5) Convective Inhibition; (6) Precipitable Water. You can then do a plot of CAPE for example as follows
import matplotlib.pyplot as plt
plt.ion()
plt.plot(time,cape,'.-')
plt.xlabel('time')
plt.ylabel('CAPE')
plt.savefig(dataPath.joinpath('cape.png'))
This is just one example, but you could plot other things in a similar way. The data I have downloaded shows that some values of CAPE are 0. These cases are not suitable for analysis. If you would like to download this file you can log in with SSH and then type the following, which will download the file to your local machine.
get SondeData/cape.png
Figure 1. Plot of the CAPE for the radio sonde data we have downloaded.
You can run the model after it has been compiled by typing
./run.sh python/namelist-mcb.in
Note that python/namelist-mcb.in is the input file for the model, and includes initial values and processes that the model will consider.
This generates a file called /tmp/<username>/output1.nc, which is a so-called NetCDF formatted file containing the model output data.
We can use Python to look at the output. There is an example script at python/example_plot_bmm.py. This can be edited to output different variables from the NetCDF file. To plot the output file at /tmp/<username>/output1.nc you can type:
python3 python/example_plot_bmm.py
This will generate a file at /tmp/<username>/Test.png
How to we obtain the output file from inside the container, so that we can view it or insert it in a report?
We can open up another CMD, or terminal window and connect with SFTP (Secure File Transfer Protocol).
From another terminal or CMD window type
sftp -i id_virtual_students.key <username>@<IP-address>
This will log you into the VM. You can bring the file over to your local system by typing
get /tmp/<username>/Test.png
And then you will be able to view it in the usual way. By default it will be transferred to the folder that you were in before you logged in with SSH.
The first calculations of how NaCl mixing ratio in the air impacts cloud albedo were done by a paper we wrote and published in Phil Trans of the Royal Society. See https://royalsocietypublishing.org/doi/full/10.1098/rsta.2014.0056
The bin-microphysics-model does not take NaCl mixing ratio as an input. It takes the aerosol lognormal parameters:
You can do a batch run using some python scripts to simulate how increasing NaCl mixing ratio in the air impacts the cloud albedo of the base case.
To run the batch run code use the batchRunsMCB.py script, which is in the python folder inside the repository. This runs 50 simulations for a range of NaCl mixing ratios, that are logarithmically spaced from /tmp/<username>. To go into the python folder enter:
cd python
Then to run the code to batch process these factors type:
python3 batchRunsMCB.py
In order to create the lookup table (LUT) we can use the lutMCB.py script. From inside the python directory, enter this command:
python3 lutMCB.py
This will create a plot in /tmp/<username>/lut.png, shown in Fig. 1. Interestingly, we see that increasing the NaCl mixing ratio in the air has a non-linear effect on cloud albedo. Why is this? At the large end the curve tends to an albedo of 1 (i.e. all the visible radiation is reflected to space).
Figure 1. A plot of the cloud albedo vs the NaCl mixing ratio for the default simulation.
You can either change the python/namelist-mcb.in file to have full control, or you can edit some of the variables that are changed in the runsDefineMCB.py file. In this file you can edit winit, logSig, and Dm, which are the vertical wind speed in m s$^{-1}$ and the size distribution parameters of the sea salt.
From the python directory, open up ipython3 by typing:
ipython3
Run batchRunsMCB.py in the following way
import runsDefineMCB
import batchRunsMCB
batchRunsMCB.batchRuns()
You can then run the lutMCB.py script by typing:
run -i lutMCB.py
As before this will generate a plot and save it in the /tmp/<username>/lut.png file.
THE STEPS BELOW CAN BE REPEATED
In a separate SSH window you can edit the runsDefineMCB.py file (make sure you are in the python folder inside the repository:
nano -l runsDefineMCB.py
You can edit Dm=20e-9 to see what effect changing the mode diameter has. Once the file is saved go back to the ipython3 window. You will need to re-import runsDefineMCB and batchRunsMCB
import importlib
importlib.reload(runsDefineMCB)
importlib.reload(batchRunsMCB)
Check that your value of Dm has been altered by typing
batchRunsMCB.Dm
It should be equal to 2e-08 if that's what you changed it to. Rerun the batchRunsMCB.batchRuns() method by typing:
batchRunsMCB.batchRuns()
when finished plot the new LUT on the existing figure
run -i lutMCB.py
This will add to the curent plot and overwrite / save it in the /tmp/<username>/lut.png file.
END REPEAT SECTION
You may want to add a legend:
plt.legend(('$D_m=100\,nm$','$D_m=20\,nm$'))
plt.savefig('/tmp/mccikpc2/lut.png')
The figure will look as in Fig 2.
Figure 2. A plot of the cloud albedo vs the NaCl mixing ratio for two different mode diameter sizes. The smaller mode size leads to a higher albedo for lower values of NaCl mixing ratio.
To understand whether implementing MCB will be of benefit we need to assess the cost vs the benefit. The paper by Connolly et al (2014) assesses the cost of the different spray techniques. These are roughly as follows
Rayleigh instability:
Taylor cone-jets:
Supercritical flow:
Effervescent spraying:
where
We want to calculate the average global surface temperature, given the values of greenhouse gasses and the albedo.
The fraction of the globe that is covered by clouds that can be used for MCB is around
The albedo of the planet,
where
NaCl aerosol must fill the atmospheric marine boundary layer, which is typically 1000 m deep. The total volume of air is therefore the fraction,
The total mass of NaCl is this volume, multiplied by the density of air,
This aerosol is thought to have a lifetime of
The salinity is the mass of salt (kg) divided by the volume of water (
This can be used to calculate the required power of all of the sprayers.
The Stern Review (2007) gives some estimates of the cost to society of rising global temperatures. Key messages from this review are:
- 2-3
$^\circ C$ warming results in a 0-3% loss in global GDP - 5-6
$^\circ C$ warming results in a 5-10% loss in global GDP - These are lower estimates. If there is an amplification, it may lead to a further 5-7% increase in the cost or 11-14% if non-market impacts are included.
- If we factor in the disporportionate burden of climate change the cost could increase by more than 25%.
Current global GDP is approx. 100 trillion USD
The gradient of percent loss in global GDP to rise in temperature is:
This means that the intercept is
So the a straight line fit to this data is:
This article hints that the cost would be between 3.2 and 4.8 billion USD to build the ships. Assume maintenance is 2.5% of that each year (a pure guess, can this estimate be improved?)
You can't argue that the energy for the sprayers will be free, even if it is derived from wave energy. The energy must have a cost. The device to convert wave energy to useful energy for the sprayers needs to be manufactured and maintained, so we can argue there will be a cost of energy per joule. The estimate used is
You will need to have run the batchRunsMCB.py script to generate the files for the LUT. Then change directory so you are in the mcb_analysis folder
cd mcb_analysis
You can then edit the script geoengineering_analysis.py to choose the spray technology. To do this set method on line 66 to one of:
- RAYLEIGH_JET
- TAYLOR_CONE
- SUPERCRITICAL
- EFFERVESCENT
Run the script with:
python3 geoengineering_analysis.py
Each time you run the script, 3 plots will be saved in /tmp/<username>/. Examples of the cost-benefit analysis are shown in Figure 3.
Figure 3. Plots of the cost-benefit analysis for Rayleigh spraying; Taylor Cones, Supercritical, and Effervescent techniques
-
Connolly P. J., McFiggans G. B., Wood R. and Tsiamis A. 2014Factors determining the most efficient spray distribution for marine cloud brighteningPhil. Trans. R. Soc. A.3722014005620140056 http://doi.org/10.1098/rsta.2014.0056
-
Levenson B. P., Planet temperatures with surface cooling parameterized, Advances in Space Research, Volume 47, Issue 11, 2011, Pages 2044-2048, ISSN 0273-1177, https://doi.org/10.1016/j.asr.2011.01.031
-
Stern, N. (2007). Economic Modelling of Climate-Change Impacts. In The Economics of Climate Change: The Stern Review (pp. 161-190). Cambridge: Cambridge University Press. doi:10.1017/CBO9780511817434.012
-
Latham, J., Bower, K., Choularton, T., Coe, H., Connolly, P., Cooper, G., Craft, T., Foster, J., Gadian, A., Galbraith, L., Iacovides, H., Johnston, D., Launder, B., Leslie, B., Meyer, J., Neukermans, A., Ormond, B., Parkes, B., Rasch, P., Rush, J., Salter, S., Stevenson, T., Wang, H., Wang, Q., Wood, R., 2012. Marine cloud brightening. Philos. Trans. A. Math. Phys. Eng. Sci. 370, 4217–62. https://doi.org/10.1098/rsta.2012.0086