Spatial Calculation of the Standardized Precipitation Evapotranspiration Index

The Standardized Precipitation Evapotranspiration Index (SPEI) offers a more accurate representation of drought conditions compared to the Standardized Precipitation Index. This is because it incorporates evapotranspiration calculations, thus making it superior to the latter. For this study, we calculated the spatial distribution of the SPEI index over a 12-month period in a particular region. The results clearly depict alternating patterns of drought (in red) and moisture (in blue) in the area. For instance, from January 2000 to March 2003, the extent of the drought gradually expanded to cover the entire region by January 2002. The severity of the drought also deepened during this period. However, afterwards, the drought started to weaken, indicating a gradual diminishing trend.

Spatial calculation of the Standardized Precipitation Index

The Standardized Precipitation Index (SPI) is one of the metrics used to assess drought conditions. Because it only requires precipitation data and is relatively simple to calculate, it has gained widespread application. In this study, we calculated the SPI over a six-month timescale for a specific location, which revealed a clear alternation between wet (blue) and dry (red) periods.

Batch interpolation of meteorological variables with ANUSPLIN

ANUSPLIN is a globally renowned software for interpolating meteorological variables and has been widely used in meteorological interpolation activities around the world. In this instance, it is used to interpolate meteorological station data to obtain the temperature distribution for a particular area from December 1964 to 2022. The interpolation process utilizes Matlab code to coordinate and manage ANUSPLIN's input, operation, and output phases.

References

[1] ANUSPLIN.

CloudComputing: How can I execute the Matlab code in CCDB

CCDB(https://ccdb.alliancecan.ca/, aka Compute Canada Database)

First, you should debug the code locally to ensure it runs smoothly. Also, make sure that all required custom functions are included in the directory you plan to upload.

Secondly, download Globus for transferring code and data.

Third, establish a connection with cedar using the built-in Command Prompt in Windows.

1 2	ssh -Y username@cedar.computecanada.ca password

Create an executable file. First, navigate to the directory where the starting code for execution is located using the "cd" command.

Below is the template for the executable file. Prepare it in advance, and after creation, simply copy and paste it in. To save, use Ctrl+O, then Enter, and then Ctrl+X.

1 2 3 4 5 6 7 8 9 10 11

#!/bin/bash #SBATCH --time=20-00:00 #SBATCH --nodes=1 #SBATCH --cpus-per-task=12 #SBATCH --account=def-XXXXX #SBATCH --mem-per-cpu=250G #SBATCH --output=job_output.log module load matlab/2023a srun matlab -nodisplay -singleCompThread -r "MAIN_SPI_RASTER"

or

1 2 3 4 5 6 7 8 9 10 11

#!/bin/bash #SBATCH --time=20-00:00 #SBATCH --gres=gpu:4 #SBATCH --cpus-per-task=12 #SBATCH --account=def-XXXX #SBATCH --mem-per-cpu=16384M #SBATCH --output=job_output.log module load matlab/2023a srun matlab -nodisplay -singleCompThread -r "MAIN_SPI_RASTER"

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19

$ module spider matlab Versions: matlab/2014a matlab/2016b matlab/2017a matlab/2018a matlab/2018b matlab/2019a matlab/2019b matlab/2020a matlab/2020b matlab/2020b.4 matlab/2020b.6 matlab/2021a.1 matlab/2021a.5 matlab/2021b.3 matlab/2022a matlab/2022b.2 matlab/2023a.3

Python

1 2 3 4 5 6 7 8 9 10

#!/bin/bash #SBATCH --time=1-00:00 #SBATCH --gres=gpu:1 #SBATCH --cpus-per-task=1 #SBATCH --account=def-XXXX #SBATCH --mem-per-cpu=6384M #SBATCH --output=job_output.log module load python/3.11.2 python cc.py

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18

$ module spider python Versions: python/2.7.14 python/2.7.18 python/3.5.4 python/3.6.3 python/3.6.10 python/3.7.0 python/3.7.4 python/3.7.7 python/3.7.9 python/3.8.0 python/3.8.2 python/3.8.10 python/3.9.6 python/3.10.2 python/3.11.2 python/3.11.5

After successfully submitting the job, you can check its status at any time using the command "sq". Once the job has completed, the system won't provide any notifications, so it's important to keep an eye on its status.

Use scancel with the job ID to cancel a job:

1	$ scancel <jobid>

1 2	/*Check the running status.*/ $ diskusage_report

Notes

Name	Storage	Max Duration	Globus endpoint
Cedar	1T	28 days	computecanada#cedar-globus
Graham	1T	7 days	computecanada#graham-globus
Beluga	1T	7 days	computecanada#beluga-dtn
Narval	1T	7 days	Compute Canada - Narval

When running on CC, never use absolute paths; always use relative paths.

The CC platform has a maximum file limit of 500K, so it's important to use resources efficiently and reduce the number of files during runtime.

1 2	% An absolute path /project/usernumber/username/Mat/SPEI/input/prec/prec_192404.tif

CC is a UNIX system and cannot directly run .EXE files. To run an .EXE file on CC, you first need to check if Wine is installed on CC. You can do this by typing 'wine' in the Command Prompt. If it returns the following, it means Wine is already installed.

wine
Usage: wine PROGRAM [ARGUMENTS...]   Run the specified program
       wine --help                   Display this help and exit
       wine --version                Output version information and exit

Check out the size of folder on Compute Canada

% Check the size of the folder
$  du -sh foldername_1  foldername_2 

Copy a folder to a specific path

% Copy the target folder to a specific folder
$ cp -r source_folder_path destination_folder_path

References

[1] Operating on Compute Canada Clusters: a brief introduction. 2020-05-07.

[2] How to run your R script with Compute Canada. 2020-10-16.

[3] How to run Python code in CC.

[4] Using Matlab effectively on Graham and Cedar. 2018-11-07

[5] Technical support: Support request example.

[6] Run an executable file (.exe) from matlab in a linux enviroment. 2011-06-22