In terms of spatial extent (wide spread, limited), I believe that earthquakes generally have the most limited spatial extent of the three, followed by hurricanes. However, I am unsure about droughts. I assume that they have the largest spatial extent of the three, but would anyone be able to give an explanation as to why?
With regard to their predictability (hazard assessment), I would say that the order (from most predictable to least) would be:
Hurricanes as past trends can be assessed and there are specific conditions that are required for a hurricane to form (e.g. between 5, 20 degrees from equator).
I am less sure about prediction methods for droughts, but I assume that precipitation levels can be monitored, etc.
Lastly, I am aware that historic trends can be assessed for earthquakes and the general areas that will be affected can be predicted, however it is difficult to predict where the earthquake will occur along the fault line and the magnitude (I am aware that there are numerous questions on the predictability of earthquakes)
I would love it if you could share your thoughts in terms of how you would order the hazards and why.
The area potentially affected by hurricanes varies on each continent, and requires global climatic modelling to define. The oceanic area in which hurricanes can develop has been discussed previously in the Earth Science stack exchange, but the single most important factor is the area with a sea surface temperature of >27 deg C. This is also a variable in these times of climate change.
The probability of a hurricane hitting the eastern United States in the next decade probably approaches 1.0 but to narrow that down to specific cities is more difficult. Once one has formed, the meteorologists are getting very good at predicting where it will make landfall. In between, the probability of a hurricane striking a particular city in a given year is low.
With regards to earthquakes, there are many known faults where measurements indicate a high probability for a large earthquake. However, the timing may be within hundreds or thousands of years. We can predict where but not when.
These are all different ways of looking at prediction that illustrate the considerations in how we think about 'prediction'. Ranking them is comparing apples to oranges.