how to plot ecdf with > 0 value in a subplot?

Hello all, I am biased correcting climate data. In Figure below subplot (a) and (b) are fine as data is less or greater than zero, however, subplot (c) due to a lot of zero values, the bias correction doesn't appear very clear. Is there anyway, I can plot only data that correspond to values greater > 0. Here is my code and the sample figure

subplot(1,3,1)
[f,x]=ecdf(BMinT_his_EC);
plot(x,f,'r','LineWidth',1)
hold on
[f,x]=ecdf(BMinT_his_RRCM5);
plot(x,f,'k','LineWidth',1)
[f,x]=ecdf(BMinT_his_CRCM5);
plot(x,f,'b','LineWidth',1)
hold off
text(-40,0.95,'$\textcircled{\bf a}$', 'Interpreter', 'latex','color','k','FontSize',20);
axis tight
ylabel('Cumulitive frequency');
xlabel('Temperature (\circC)')

subplot(1,3,2)
[f,x]=ecdf(BMaxT_his_EC);
plot(x,f,'r','LineWidth',1)
hold on
[f,x]=ecdf(BMaxT_his_RRCM5);
plot(x,f,'k','LineWidth',1)
[f,x]=ecdf(BMaxT_his_CRCM5);
plot(x,f,'b','LineWidth',1)
hold off
text(-40,0.95,'$\textcircled{\bf b}$', 'Interpreter', 'latex','color','k','FontSize',20);
axis tight
ylabel('Cumulitive frequency');
xlabel('Temperature (\circC)')

subplot(1,3,3)
[f,x]=ecdf(BPrec_his_EC);
plot(x,f,'r','LineWidth',1)
hold on
[f,x]=ecdf(BPrec_his_RRCM5);
plot(x,f,'k','LineWidth',1)
[f,x]=ecdf(BPrec_his_CRCM5);
plot(x,f,'b','LineWidth',1)
hold off
text(60,0.95,'$\textcircled{\bf c}$', 'Interpreter', 'latex','color','k','FontSize',20);
ylabel('Cumulitive frequency');
xlabel('Precipitation (mm)')

set(gcf,'NextPlot','add');
axes; 
set(gca,'Visible','off'); 
h = title('Bias correction of Regional Climate Modem (RCM) ver 4','fontweight','b','fontSize',16);
set(h,'Visible','on');

Answer 1

[the cyclist]

You can manipulate the data to get what you want to do, but I think it is easier to just zoom in on the relevant part of the plot. How about just doing

set(gca,'YLim',[0.7 1])

[Hydro]

Like this one too. Many thanks

Answer 2

[Cam Salzberger]

You can use logical indexing to do this easily:

whichVals = x > 0;
plot(x(whichVals), f(whichVals), ...)

-Cam

[the cyclist]

I think this approach will actually change the cumulative distribution values, and therefore probably not what OP wants.

rng default

N = 1000;
NZ =1000;
t = [rand(N,1); rand(NZ,1)/100];

% Original
figure
ecdf(t)

% Zoom
figure
ecdf(t)
set(gca,'YLim',[0.5 1])

% Logical indexing -- note the ecdf is actually different, so not sure this is OK
figure
ecdf(t(t>0.05))

[Hydro]

excellent ideas. How about removing x from Cam's suggestion?

whichVals > 0;
plot(x(whichVals), f(whichVals), ...)

[the cyclist]

Sorry, I misread Cam's code completely. I thought he was doing the ecdf with the modified BPrec_his_EC values. His approach is perfect.

[Hydro]

Many thanks to both of you. I ended up using cyclist solution.