The Climate Data Toolbox for MATLAB - Analyzing Trends & Global Warming

This tutorial describes how to analyze trends in sea surface temperature using the Climate Data Toolbox for MATLAB (CDT).

Get CDT here: https://github.com/chadagreene/CDT

Greene, C.A., Thirumalai, K., Kearney, K.A., Delgado, J.M., Schwanghart, W., Wolfenbarger, N.S., Thyng, K.M., Gwyther, D.E., Gardner, A.S. and Blankenship, D.D., 2019. The climate data toolbox for MATLAB. Geochemistry, Geophysics, Geosystems, 20(7), pp.3774-3781. https://doi.org/10.1029/2019GC008392

Below you'll find the annotated code from today's show.

% View CDT documentation:
cdt deseason

% Load example data:
load pacific_sst

% Inspect the the variables in the workspace:
whos

% Calculate a map of mean sea surface temps:
sst_mean = mean(sst,3);

% Plot mean sea surface temperature:
imagescn(lon,lat,sst_mean)
xlabel longitude
ylabel latitude
colorbar

% View cmocean colormap options:
cmocean

% Set the thermal colormap:
cmocean thermal

% Add geopolitical borders:
hold on
borders

% Get the indices of the grid cell closest to (23N,115W):
row = near1(lat,23)
lat(row)
col = near1(lon,-115)
lon(col)

% Plot a circle at the location of interest:
plot(lon(col),lat(row),'ko')

% Get a 1D array of temperatures at the location of interest:
sst1 = sst(row,col,:);
size(sst1)

% Turn sst1 into a column vector:
sst1 = squeeze(sst1);
size(sst1)

% Clear the map figure:
clf

% Plot the time series of sst1:
plot(t,sst1)
ylabel temperature
xlabel time

% Format the date axis labels:  
datetick

% Remove the seasonal cycle from sst1:
sst1_ds = deseason(sst1,t);

% Plot the de-seasoned sst1 time series:
hold on
plot(t,sst_ds)

% Plot the linear trend in sea surface temperature:
polyplot(t,sst1)

% Plot the linear trend of deseasoned temperature:
polyplot(t,sst1_ds)

% Calculate the SST trend at the location of interest:
trend(sst1,12)

% Verify the total amount of warming that has occurred since 1950:
trend(sst1,12)*67

% Calculate a map of trends in sea surface temperature:
sst_tr = trend(sst,12);

% Clear the current figure:
clf

% Map the trends in sea surface temperature:
imagescn(lon,lat,sst_tr)
cb = colorbar;
ylabel(cb,'sst trend \circC/yr')

% Set the colormap:
cmocean('balance','pivot')

% Calculate trends with statistical p values:
[sst_tr,p] = trend(sst,12);

% Determine if the trend at the single location of interest is significant:
mann_kendall(sst1)

% Verify that random numbers do not produce a significant trend:
mann_kendall(rand(size(sst1)))

% Calculate a map of statistical significance:
mk = mann_kendall(sst);

% Add stippling where trends are significant:
hold
stipple(lon,lat,mk)

Content created by Chad A. Greene of NASA Jet Propulsion Laboratory.