105. US southern states - summary of BE temperature adjustments

In my previous post I summarized the temperature trends since 1900 of the six US states closest to the Gulf of Mexico (Texas, Louisiana, Mississippi, Alabama, Georgia and Florida). All the trends were constructed using data from the longest available temperature records in the state, all involved averaging the temperature anomalies from over 90 different station records, and none exhibited a significant positive warming trend.

Yet in every case the official Berkeley Earth (BE) trend does exhibit warming, often lots of it. The difference of course is largely down to the adjustments that Berkeley Earth make to the data via homogenization, Kriging, gridding and of course breakpoint alignment. In the post for each state (the links are here: Texas, Louisiana, Mississippi, Alabama, Georgia and Florida) I have quantified the magnitude of these adjustments, but I thought it would also be instructive to summarize them in one post just so that their full impact can be seen and compared.

The adjustments shown in the graphs below are of two types. The orange curve is the mean adjustment each month solely from breakpoint adjustments while the blue curve is the mean adjustment relative to unadjusted data from all sources of correction. This will also include homogenization, Kriging and gridding in addition to breakpoints, but it will also be affected by any difference in the chosen period for calculating the monthly reference temperatures (MRTs). The last of these will, however, only change the offset of the blue curve in the vertical direction relative to the orange one, not its slope or total change over time.

The graphs below indicate that the BE adjustments to the temperature data add between 0.5°C and 1.2°C to the final BE temperature trends. Given that we are constantly being told by climate scientists that the total global warming experienced so far is about 1.2°C, I would suggest that this is a bit of a problem.

Fig. 105.1: The Berkeley Earth (BE) temperature adjustments for Texas since 1900. The linear best fit (red line) to these adjustments for the period 1911-2010 has a positive gradient of +0.568 ± 0.003 °C per century.

Fig. 105.2: The Berkeley Earth (BE) temperature adjustments for Louisiana since 1900. The linear best fit (red line) to these adjustments for the period 1911-2010 has a positive gradient of +0.731 ± 0.004 °C per century.

Fig. 105.3: The Berkeley Earth (BE) temperature adjustments for Mississippi since 1900. The linear best fit (red line) to these adjustments for the period 1931-2010 has a positive gradient of +1.300 ± 0.007 °C per century.

Fig. 105.4: The Berkeley Earth (BE) temperature adjustments for Alabama since 1900. The linear best fit (red line) to these adjustments for the period 1931-2010 has a positive gradient of +1.231 ± 0.012 °C per century.

Fig. 105.5: The Berkeley Earth (BE) temperature adjustments for Georgia since 1900. The linear best fit (red line) to these adjustments for the period 1911-2010 has a positive gradient of +1.087 ± 0.006 °C per century.

Fig. 105.6: The Berkeley Earth (BE) temperature adjustments for Florida since 1900. The linear best fit (red line) to these adjustments for the period 1941-2010 has a positive gradient of +0.611 ± 0.010 °C per century.

104. US southern states - summary of temperature trends

Over the last month I have examined the temperature trends of five different US states (Louisiana, Mississippi, Alabama, Georgia and Florida) that surround, or are within 100km of (in the case of Georgia), the Gulf of Mexico. These all appear to have similar trends to that of Texas that I examined in Post 52. All have negative or stable temperature trends over the last 100 years. For comparison their temperature trends are republished here with identical data ranges (from 1900) and fitting ranges (1911-2010). What is clear is that none of these trends is remotely similar to either the Berkeley Earth (BE) versions for each state based on adjusted data, or the global trends published by NOAA, NASA-GISS, BE, HadCRU etc.

Fig. 104.1: The mean temperature change for Texas. The best fit has a slight negative gradient of -0.15 ± 0.15 °C per century.

Fig. 104.2: The mean temperature change for Louisiana. The best fit has a negative gradient of -0.38 ± 0.15 °C per century.

Fig. 104.3: The mean temperature change for Mississippi. The best fit has a negative gradient of -0.76 ± 0.17 °C per century.

Fig. 104.4: The mean temperature change for Alabama. The best fit has a negative gradient of -0.72 ± 0.17 °C per century.

Fig. 104.5: The mean temperature change for Georgia. The best fit has a negative gradient of -0.76 ± 0.16 °C per century.

Fig. 104.6: The mean temperature change for Texas. The best fit has a slight positive gradient of +0.08 ± 0.13 °C per century.

103. Florida - temperature trends STABLE

Like all the other US states that border the Gulf of Mexico, Florida has not experienced any global warming, but neither has the climate cooled. In fact over the last 100 years the climate of Florida has remained fairly stable as shown by the mean temperature anomaly (MTA) data for the state illustrated in Fig. 103.1 below.

Fig. 103.1: The mean temperature change for Florida relative to the 1951-1980 monthly averages. The best fit is applied to the monthly mean data from 1911 to 2010 and has a slight positive gradient of +0.08 ± 0.13 °C per century.

The MTA in Fig. 103.1 was calculated by averaging the temperature anomalies from the 100 longest temperature records for the state. All these records had over 700 months of temperature data before the end of 2013 and 31 were long stations that each had more than 1200 months of data in total. For a full list of stations see here.

The anomalies for each station were determined using the usual method as outlined in Post 47. This involved first calculating the monthly reference temperatures (MRTs) for each station using a set reference period, in this case from 1951 to 1980, and then subtracting the MRTs from the raw temperature data to deliver the anomalies. If a station had at least twelve valid temperatures within the MRT interval then its anomalies were included in the MTA calculation. In total 93 stations were included with seven being excluded for lack of data between 1951 and 1980. The total number of stations included in the MTA in Fig. 103.1 each month is indicated in Fig. 103.2 below. The peak just around 1960 suggests that the 1951-1980 interval was indeed the most appropriate.

Fig. 103.2: The number of station records included each month in the mean temperature anomaly (MTA) trend for Florida in Fig. 103.1.

The locations of the one hundred stations is shown in the map in Fig. 103.3 below. This appears to show that the geographical spread is fairly uniform, although there does appear to be a greater concentration of stations in the more highly populated areas of Tampa and Miami. These areas also appear to have more stations with warming trends where a warming trend is defined as one where the temperature gradient for 1911-2010 is positive and exceeds twice the error. Nevertheless, the variation in station density is probably not sufficient to significantly distort the average in Fig. 103.1 from its true value. In which case the simple average of the anomalies from all stations used to construct the MTA in Fig. 103.1 should still yield a fairly accurate temperature trend for the state as a whole.

Fig. 103.3: The (approximate) locations of the 100 longest weather station records in Florida. Those stations with a high warming trend between 1911 and 2010 are marked in red while those with a cooling or stable trend are marked in blue. Those denoted with squares are long stations with over 1200 months of data, while diamonds denote medium stations with more than 480 months of data.

Next I calculate the corresponding MTA result based on data that has been adjusted by Berkeley Earth (BE). The result is shown in Fig. 103.4 below.

Fig. 103.4: Temperature trends for Florida based on Berkeley Earth adjusted data from the 100 longest station data records. The best fit linear trend line (in red) is for the period 1911-2010 and has a gradient of +0.52 ± 0.05°C/century.

Comparing the curves in Fig. 103.4 with the published Berkeley Earth (BE) version in Fig. 103.5 below indicates remarkably good agreement at least as far back as 1900. This indicates that the 100 longest records are sufficient to determine the MTA for this period, and that simple averaging of anomalies is also highly effective and accurate.

Fig. 103.5: The temperature trend for Florida since 1750 according to Berkeley Earth.

The differences between the MTA in Fig. 103.1 and the BE versions using adjusted data in Fig. 103.4 and Fig. 103.5 are mainly due to the data processing procedures used by Berkeley Earth. These include homogenization, gridding, Kriging and most significantly breakpoint adjustments. These lead to changes to the original temperature data, the magnitude of these adjustments being the difference in the MTA values seen in Fig. 103.1 and Fig. 103.4. The magnitudes of these adjustments are shown graphically in Fig. 103.6 below. The blue curve is the difference in MTA values between adjusted (Fig. 103.4) and unadjusted data (Fig. 103.1), while the orange curve is the contribution to those adjustments arising solely from breakpoint adjustments. Both are considerable with the former leading to an additional warming since 1935 of over 0.5°C.

Fig. 103.6: The contribution of Berkeley Earth (BE) adjustments to the anomaly data in Fig. 103.4 after smoothing with a 12-month moving average. The blue curve represents the total BE adjustments including those from homogenization. The linear best fit (red line) to these adjustments for the period 1911-2010 has a positive gradient of +0.414 ± 0.007 °C per century. The orange curve shows the contribution just from breakpoint adjustments.

Finally there is the question of the negative discontinuity in the temperature data in 1957 that was observed in the data for Texas (Post 98), Louisiana (Post 98), Mississippi (Post 99) and Alabama (Post 101). Surprisingly, this is not obvious in the Florida data. This may suggest that the discontinuity is a natural regional phenomenon rather than man-made due to changes in national data recording protocols.

Summary 

According to the raw unadjusted temperature data, over the past century the climate of Florida has remained stable (see Fig. 103.1). Most of the state has cooled but the more populated areas around Miami and Tampa have exhibited some warming that has compensated for the cooling (see Fig. 103.3).

Over the same period adjusted temperature data from Berkeley Earth claims to show that the climate of Florida has warmed by as much as 0.5°C (see Fig. 103.4 and Fig. 103.5).

Acronyms 

BE = Berkeley Earth.

MRT = monthly reference temperature (see Post 47).

MTA = mean temperature anomaly.

Link to list of all stations and their raw data files.