Yes, I see there was a problem with the structure of the caption. I have fixed it. Basically, each map shows the result of subtracting the out the typical El Niño influence. Subtracting out the typical El Niño influence from the observations shows us the "competing influences" that shaped the precip pattern. We see that there is a dry signal over the Southwest and Mexico.

The question for forecasters is whether that dry influence was "internal variability" (i.e. chaotic, unpredictable at the start of the season) or if it might reflect some other seasonal influence besides El Niño that might have been predictable in advance. To start to answer that question, we can look at what the forecast would have looked like if we subtract out the typical El Niño influence. That is what the left hand map shows. Here, we also see the same dry signal over the Southwest and Mexico that we see in the observations. This suggests that before the season began, the models picked up on something operating in the climate system that favored dry conditions over that area, something predictable at the start of the season that was working against the wet influence that El Niño typically has in that area.