Note: updated graphs are now posted on the MLMW page accessible in the menu bar at the top of this page.
Mauna Loa on the big island of Hawaii is probably more famous for being a large active shield volcano and for carbon dioxide and solar measurements than for temperature measurements. However, the Mauna Loa Observatory (MLO) has been making continuous temperature measurements since 1977 and the data return is overall very good. The observatory is run by the Earth System Research Laboratory (ESRL) which is part of the US National Oceanic and Atmospheric Administration (NOAA). For more information, see this link: Mauna Loa Observatory.
The observatory is at an elevation of 3,397 meters (11,141 feet) above mean sea level about 5 kilometers (3 miles) north of the Mauna Loa summit that reaches to 4,169 meters (13,679 feet) above sea level. See the map below for the position of the observatory relative to the summit.
Below is a photograph of the MLO looking north toward Mauna Kea. The tall tower has temperature sensors at 2 meters, 10 meters, and 35 meters above ground level.
I downloaded hourly meteorological data that is available by FTP from the ESRL web site here: FTP Data Finder. I loaded the data into an Excel workbook by year and compiled daily, monthly, and annual statistics for temperature at 2 meters above ground level. The average annual data return for temperature at 2 meters over the period from 1977 through 2015 was 97%, but there were a few fairly large gaps in the data. The largest gap was from March 29 to April 29 of 1984. Since most of April was missing, I removed it from the data set. I also compiled annual weighted means using the monthly averages in an attempt to make sure every month was weighted the same from year to year even if there was substantial missing data. This is not a perfect approach and is subject to increased uncertainty if there is a large data gap near the beginning or end of a month where temperatures are normally rising or falling on average over the course of the month. So, the incompleteness does add a small amount of uncertainty to the results, especially for the years with the lowest data returns – 1984, 1993, and 2001 with data completeness of 87, 88, and 88 percent respectively.
The annual average temperatures and associated trend are displayed in Figure 1. The data show a fairly significant upward trend of +0.0264 degrees Celsius (C) per year (Coefficient of Determination R2=0.2537). This temperature trend corresponds to +1.03C over the 39 year period of record and to +2.64C if it continued for 100 years.
I noticed that since the high peak in 1995, the rise looks much weaker than before the peak. So I made separate graphs for 1977-1994 shown in Figure 2 and for 1995-2015 shown in Figure 3. The temperature trends in these graphs confirm that the rise was indeed steeper before 1995 than since 1995.
The temperature trend for 1977 through 1994 was +0.0345C per year, corresponding to -0.62C over the 18 year period or to +3.45C if that trend continued for 100 years. The temperature trend for 1995 through 2015 was +0.194C per year, which is +0.41C over the 21 year period and would be only +1.94C if it continued for 100 years. The statistical confidence in both of these trends is weak with R2=0.1038 for 1977-1994 and R2=0.0539 for 1995 to 2015.
Even though MLO is in the tropics at a latitude of 19.5 degrees north of the equator, the temperature shows a pronounce seasonal pattern as can be seen in Figure 4. On average the highest monthly average temperature is for June and the lowest for February.
The most recent 1981-2010 standard climatological reference period averages are 5.2C for February and 9.1C for June. Figures 5 and 6 show the temperature trend graphs for these two months respectively.
The trend for February is only +0.0106C per year, corresponding to +0.41C over the 39 year period and to +1.06C if it continued 100 years. The trend for June is much higher at +0.0279C per year, or +1.09C over the 39 year period and would be +2.79C if it continued 100 years. The February graph also shows a larger variability in monthly average temperature from year to year compared to the June graph.
Regardless how the data are sliced, the temperature trend at the MLO is significantly upward, although with a hint of a slower increase since 1995. There is a slight chance that expansion of the observatory over time could have an effect similar to a miniature urban heat island that increases over time with the expansion. How much influence the expansion might have on the temperature trend is difficult to determine without additional information about potential heat sources at the observatory and how they may have changed over time. Even if we assume that such influences are insignificant, the upward temperature trend could be from natural causes and not necessarily from CO2. From what I recall reading, hypothetical warming induced by CO2 should occur primarily at higher latitudes and not in the tropics. Consequently, the observed upward trend is likely to be from unknown natural causes or possibly from increasing localized heat influence over time or both.