How to Plot Time Series Data of a Temperature Map

Hi,

I am trying to plot a timeseries data of the SST at certain locations on the world but I keep getting a blank plot. This is my code:

load global_sst
figure
imagescn(lon,lat,mean(sst,3))
hold on

gomlon = [21.25  1.25  -18.75 111.25  -116.25  176.25 -91.25 136.25];
gomlat = [26.25  11.25  -13.75   61.25   16.25   31.25 -76.25 -23.75];
plot(gomlon,gomlat,'ro')

text(21.25 ,26.25,'1','color','red','horiz','center','vert','bot');
text(1.25 ,11.25,'2','color','red','horiz','center','vert','bot');
text(-18.75 ,-13.75,'3','color','red','horiz','center','vert','bot');
text(111.25 ,61.25,'4','color','red','horiz','center','vert','bot');
text(-116.25 ,16.25,'5','color','red','horiz','center','vert','bot');
text(176.25 ,31.25,'6','color','red','horiz','center','vert','bot');
text(-91.25 ,-76.25,'7','color','red','horiz','center','vert','bot');
text(136.25 ,-23.75,'8','color','red','horiz','center','vert','bot');

%% Location 1
gomlon1 = [18.75 18.75 23.75 23.75];
gomlat1 = [28.75 23.75 28.75 23.75];

[Lon,Lat] = meshgrid(lon,lat);
mask1 = geomask(Lat,Lon,gomlat1,gomlon1);

contour(Lon,Lat,double(mask1),[0.5 0.5],'b')
%% Location 2
gomlon2 = [3.75 3.75 -1.25 -1.25];
gomlat2 = [13.75 8.75 13.75 8.75];

[Lon,Lat] = meshgrid(lon,lat);
mask2 = geomask(Lat,Lon,gomlat2,gomlon2);

contour(Lon,Lat,double(mask2),[0.5 0.5],'b')
%% Location 3
gomlon3 = [-21.25 -21.25 -16.25 -16.25];
gomlat3 = [-16.25 -11.25 -16.25 -11.25];

[Lon,Lat] = meshgrid(lon,lat);
mask3 = geomask(Lat,Lon,gomlat3,gomlon3);

contour(Lon,Lat,double(mask3),[0.5 0.5],'b')
%% Location 4
gomlon4 = [113.75 113.75 108.75 108.75];
gomlat4 = [58.75 63.75 58.75 63.75];

[Lon,Lat] = meshgrid(lon,lat);
mask4 = geomask(Lat,Lon,gomlat4,gomlon4);

contour(Lon,Lat,double(mask4),[0.5 0.5],'b')
%% Location 5
gomlon5 = [-118.75 -118.75 -113.75 -113.75];
gomlat5 = [18.75 13.75 18.75 13.75];

[Lon,Lat] = meshgrid(lon,lat);
mask5 = geomask(Lat,Lon,gomlat5,gomlon5);

contour(Lon,Lat,double(mask5),[0.5 0.5],'b')
%% Location 6
gomlon6 = [178.75 173.75 178.75 173.75];
gomlat6 = [33.75 28.75 33.75 28.75];

[Lon,Lat] = meshgrid(lon,lat);
mask6 = geomask(Lat,Lon,gomlat6,gomlon6);

contour(Lon,Lat,double(mask6),[0.5 0.5],'b')
%% Location 7
gomlon7 = [-93.75 -93.75 -88.75 -88.75];
gomlat7 = [-78.75 -73.75 -78.75 -73.75];

[Lon,Lat] = meshgrid(lon,lat);
mask7 = geomask(Lat,Lon,gomlat7,gomlon7);

contour(Lon,Lat,double(mask7),[0.5 0.5],'b')
%% Location 8
gomlon8 = [138.75 138.75 133.75 133.75];
gomlat8 = [-26.25 -21.25 -26.25 -21.25];

[Lon,Lat] = meshgrid(lon,lat);
mask8 = geomask(Lat,Lon,gomlat8,gomlon8);

contour(Lon,Lat,double(mask8),[0.5 0.5],'b')
%%
figure
sst_gom = local(sst,mask7,'omitnan');
t=1:736;
figure(2)
plot(t,sst_gom)
axis tight
datetick('x','keeplimits')
ylabel 'y'

As an example, I tried plotting the time series plot of location 7. I did this by making a mask of the location, then tried plotting the data within this mask. However I get a blank plot. Any advice?

Also as a note, I am using the Climate Data Toolbox by Chad A. Greene and tried following his example.

Thanks.

Answer 1

[Chad Greene]

Hi John,

A couple of things:

1. Location number 7 is in Antarctica! The sea surface temperature data is all NaN on land, so averaging the grid cell values around Location 7 will just give you NaN.
2. The global_sst.mat dataset is a static average, not a time series. This means that even in the ocean where all SST values are valid, if you take the average SST in a given mask, you'll just end up with a single scalar value. If you would like to apply the local function to an example time series of SSTs, use the pacific_sst.mat dataset, which contains a cube where the first two dimensions are spatial, and the third dimension is time.

[John Ziggs]

Hi Chad,

Thank you for your response. Your explanation makes sense. I guess my next question is what if the matrice dataset is not in the two dimension spatial, third dimension time format? I have to map out and find the time series of the dataset attached and it consists only of temperature measurements.

I mapped it out using:

inputdirTS = 'C:\Users\john\Desktop\HW 4\TS'; % directory where files are located

S = dir(fullfile(inputdirTS,'*.txt')); 
for k = 1:numel(S) 
    fnm = fullfile(inputdirTS,S(k).name);
    vectorTS = load(fnm);
    TS = flipud(reshape(vectorTS,[72 144])); %diurnal (daily) temperature variations
end

figure(1)
[lat,lon] = cdtgrid(2.5); % creates a 2.5 degree global grid
pcolor(lon,lat,TS) % plots the grid
shading interp 
cmocean thermal % sets the colormap
borders('countries','color',rgb('gray'))
xlabel longitude
ylabel latitude 
title('Daily Average Sea Surface Temperature (TS)')

Then I did the geomask for one location and it loaded really slow.

gomlon = [21.25  1.25  -18.75 111.25  -116.25  176.25 -91.25 136.25];
gomlat = [26.25  11.25  -13.75   61.25   16.25   31.25 -76.25 -23.75];
plot(gomlon,gomlat,'ro')

text(21.25 ,26.25,'1','color','red','horiz','center','vert','bot'); % Location 1
text(1.25 ,11.25,'2','color','red','horiz','center','vert','bot');
text(-18.75 ,-13.75,'3','color','red','horiz','center','vert','bot');
text(111.25 ,61.25,'4','color','red','horiz','center','vert','bot');
text(-116.25 ,16.25,'5','color','red','horiz','center','vert','bot');
text(176.25 ,31.25,'6','color','red','horiz','center','vert','bot');
text(-91.25 ,-76.25,'7','color','red','horiz','center','vert','bot');
text(136.25 ,-23.75,'8','color','red','horiz','center','vert','bot');

%% Location 1
gomlon1 = [18.75 18.75 23.75 23.75]; %Location 1 lon
gomlat1 = [28.75 23.75 28.75 23.75]; % Location 1 lat

[Lon,Lat] = meshgrid(lon,lat);
mask1 = geomask(Lat,Lon,gomlat1,gomlon1);

contour(Lon,Lat,double(mask1),[0.5 0.5],'b')

Moving on from here, I'm not sure how I would go about doing a time series plot of the mask of Location 1. I tried this syntax but this is for the entire globe rather than Location 1.

figure
TS = mask1(:);
TS = TS(TS > 0);
x = timeseries(vectorTS);
% x = timeseries(TS); % plot gave a straight line
x.Name = 'Daily Count';
x.TimeInfo.Units = 'days';
x.TimeInfo.StartDate = '04-06-01 00:00';     % Set start date.
%x.TimeInfo.End = '04-08-31 21:00';
x.TimeInfo.Units = 'hours';
x.TimeInfo.Increment = 3;
x.TimeInfo.Format = 'yy-mm-dd HH:MM';       % Set format for display on x-axis.
x.Time = x.Time - x.Time(1);        % Express time relative to the start date.
plot(x)

Any help and advice helps.

Thanks,

John