Help with ocean transect plot

Hi Matlabbers,

I'm fairly new to the community and struggling to make a transect plot of some nitrate data that I have from CTD casts. I have 11 stations across an area, each with different depths, and each with various nitrate concentration data.

So, what I want to produce is a transect plot, showing water depth on the y axis, distance between stations on x-axis and the nitrate concentration at each station on the z-axis, but in a 2D plot.

My issue is, for each station, the depth varies, and my nitrate concentration data is quite scattered, hence I want to interpolate it so that for missing data (NaNs), it takes the concentration from neighbouring stations at the same depth and generates a value. When I do this using my data, Matlab interpolates across the rows which doesnt work since there are various depths across each row. I can't really make a uniform depth profile because it differs so much between each station. I've tried fillmissing and interp1 to get around the fact that i have so many NaN values, but the results are really blocky.

Any suggestions?

This is what I currently get:

And this is the sort of look I'm after

Thanks in advance :)

Answer 1

[Ruchika Parag]

Hi Lisa,

Here are some suggestions to help you achieve your goal:

1. Interpolate Data in 2D: Use griddata or scatteredInterpolant to interpolate your data in both depth and distance dimensions. This will help in filling NaN values more effectively.
2. Create a Regular Grid: Generate a regular grid for both depth and distance to interpolate your scattered data onto this grid.
3. Plotting: Use pcolor or contourf to plot the interpolated data.

Here is a code snippet explaining how you can do this:

% Define stations used
stn_use_SC = [5 8 9 14 15 16 19 20 21 22 23];
stn_name_SC = {'SC2' 'SC3' 'SC4' 'SC5' 'SC6' 'SC7' 'SC8' 'SC9' 'SC10' 'SC11' 'SC12'};

% Cumulative distance between each station
cumdistS = [0 2.6087 3.5718 4.7268 5.3693 5.9175 6.5523 7.1430 7.3480 7.8354 7.9372];

% Extract relevant data
depths = Depth(:, stn_use_SC);
nitrate = Nitrate(:, stn_use_SC);

% Flatten the data to create vectors for griddata
depths_vector = depths(:);
cumdistS_vector = repmat(cumdistS, size(depths, 1), 1);
cumdistS_vector = cumdistS_vector(:);
nitrate_vector = nitrate(:);

% Remove NaN values from the vectors
valid_indices = ~isnan(nitrate_vector);
depths_vector = depths_vector(valid_indices);
cumdistS_vector = cumdistS_vector(valid_indices);
nitrate_vector = nitrate_vector(valid_indices);

% Create a regular grid for interpolation
depth_grid = linspace(min(depths_vector), max(depths_vector), 100);
distance_grid = linspace(min(cumdistS_vector), max(cumdistS_vector), 100);
[distance_mesh, depth_mesh] = meshgrid(distance_grid, depth_grid);

% Interpolate nitrate data onto the regular grid
nitrate_interpolated = griddata(cumdistS_vector, depths_vector, nitrate_vector, distance_mesh, depth_mesh, 'linear');

% Plot the interpolated data
figure;
pcolor(distance_mesh, depth_mesh, nitrate_interpolated);
shading interp;
colorbar;
colormap('parula');
hold on;

% Plot the original data points
plot(cumdistS, depths, 'k.', 'MarkerSize', 10);

% Add contour lines
contour(distance_mesh, depth_mesh, nitrate_interpolated, 'k', 'ShowText', 'on');

% Adjust the position of station names to be above the plot area
y_label_position = -max(depth_grid) * 0.05; % Adjust this value as needed
for i = 1:length(stn_name_SC)
    text(cumdistS(i), y_label_position, stn_name_SC{i}, 'fontsize', 5, 'rotation', 45, 'HorizontalAlignment', 'right');
end

% Label axes
xlabel('Distance (km)');
ylabel('Depth (m)');

% Reverse both axes
set(gca, 'XDir', 'reverse', 'YDir', 'reverse', 'fontsize', 12);

% Adjust y-axis limits
ylim([0 400]);

[Lisa Friberg]

Amazing, that worked perfectly! Thank you so so much!