The Extended Outlook webpage provides the probability forecasts of different weather elements as well as their evolution trends in the next 14 days. The above diagram shows the probability forecast of daily minimum temperature with its evolution shown as a time series.

Updated once a day at around noon, the minimum temperature probability forecast is automatically generated by computer using data from ensemble prediction systems of weather models. The 9-Day Weather Forecast issued by forecasters is based on various meteorological observations, computer forecast model products including the probability temperature forecast and past forecast experience. As the two forecast methods are different, there could be differences between the two forecasts. Please click here for the detailed 9-Day Forecast covered by the gray area.

The colours in the graph represent the chances of minimum temperature occurring in the marked ranges. In general, reddish colours for higher chance, yellow for medium chance (around 50%) and bluish for lower chance. For instance, if the colour of forecast minimum temperature for the range 23-25 degrees is orange at a certain forecast time period, it means that there will be a 60-70% chance for minimum temperature to occur in that range by that time. Also shown on the graph is one solid gray line sandwiched between an upper and a lower dotted gray line. The three gray lines together denote the climatological distribution, namely the mean, 85^th and 15^th percentiles respectively. That is, 70% of the actual pressure values registered in the same period over past years fall within the range enclosed by the upper and lower dotted lines.

According to verification results, the average difference between the most probable minimum temperature forecast and the actual records at the Observatory generally increases with forecast time and is about 1-2 degrees.