Michael,
A workaround would be to underlay the FILL plot with a SHADE. The shade command does fill to the edges of each grid cell. If you use /LEVELS (with no argument to repeat the color levels from the previous plot), you will get the desired colors.
SHADE/lin/key/t=1-dec-1599:1-feb-1601/hl=1-jan-1600:1-jan-1601 t_ref[d=1,x=@ave]
FILL/OVER/NOLAB/LEV/lin/key/t=1-dec-1599:1-feb-1601/hl=1-jan-1600:1-jan-1601 t_ref[d=1,x=@ave]
We will still work on tracking down the bug in FILL.
Ansley
On 3/7/2016 8:43 AM, Michael Erb wrote:
MichaelI want to use the "fill" command to make a Hovmoller plot over the course of the year for a climatological variable. By default, ferret doesn't interpolate over the end-of-year break, leaving white space for the first half of January and the last half of December when using monthly data. This doesn't look great for a published figure. Is there a good way to ensure that data is plotted there?
One thing that I tried was to define the timespan and the hlimit to get ferret to plot that time of the year as well (e.g. fill/lin/key/t=1-dec-1599:1-feb-1601/hl=1-jan-1600:1-jan-1601 t_ref[d=1,x=@ave]). This fixes my first problem, but introduces a new one: all latitudes where the color of the plot doesn't change over the course of the year are plotted white. I assume the problem is because the left and right "edges" of the data are outside of the plot, but I'm not sure how to fix it. I've attached a figure to show what I mean. The plot on the left is default ferret behavior. The plot on the right is my attempt to fix it, but with the horizontal white bands.
Is there a better way to do what I'm trying to do?