132: UHI #5 - Pretoria (South Africa)

The three largest cities in southern Africa are Kinshasa, Johannesburg and Nairobi. All have populations of more than ten million, so all three could be good contenders as examples of the urban heat island (UHI) effect. Unfortunately in all three cases making a definitive assessment is difficult because these cities do not have data of high enough quality.

The city in southern Africa with the next highest population is Luanda in Angola with a population of eight million people. Luanda does have good temperature data stretching back to 1879 that does appear to show a strong warming trend even though the data after 1980 is fragmented, probably due to the civil war. The problem is that there is very little other good temperature data for Angola (see here for a complete list of stations), so there is no reliable trend for Angola as a region as I showed in Post 82, and so no accurate regional trend with which to compare the Luanda data.

The country in southern Africa with the the best temperature data is South Africa, and while Johannesburg has no high quality weather stations near its centre, the city of Pretoria (which is part of the same conurbation) does, although the temperature record for Pretoria Eendracht (Berkeley Earth ID: 159076) only starts in 1949. Nevertheless, since then the respective temperature trends show that Pretoria has warmed significantly more than South Africa as a whole (see Fig. 132.1 below) with up to 3°C of warming in Pretoria but less than 1°C in South Africa.

Fig. 132.1: The change to the 5-year average temperatures of Pretoria Eendracht (red curve) and South Africa (blue curve) since 1952.

In Post 37 I examined the temperature trends for South Africa. The mean temperature change since 1880 is shown in Fig. 132.2 below and it indicates that South Africa exhibited no significant warming before 1980 but has since warmed by about 0.7°C. In fact the best fit for 1951-2010 indicates a temperature rise of about 1.08°C in 60 years while the 5-year average suggests a rise of about 0.96°C.

Fig. 132.2: The mean temperature change for South Africa since 1857 relative to the 1961-1990 monthly averages. The best fit is applied to the monthly mean data from 1951 to 2010 and has a positive gradient of +1.80 ± 0.14 °C per century.

In contrast to the rest of South Africa, Pretoria Eendracht (Berkeley Earth ID: 159076) shows significant and continuous warming since 1950 (see Fig. 132.3 below). The best fit for 1951-2010 indicates a temperature rise of more than 2.81°C in 60 years while the 5-year average suggests a rise of 2.99°C.

Fig. 132.3: The mean temperature change for Pretoria Eendracht since 1949 relative to its 1961-1990 monthly averages. The best fit is applied to the monthly mean data from 1951 to 2010 and has a positive gradient of +4.69 ± 0.24 °C per century.

Summary

The following temperature changes were observed from 1951 to 2010.

South Africa: 0.96°C (trend 1.08°C).

Pretoria: 2.99°C (trend 2.81°C).

So Pretoria has warmed by at about 2°C more than the surrounding state of South Africa, or up to three times faster. A classic UHI!

117: Cameroon - temperature trends STABLE before 1990

Like Chad (see Post 15) and the Central African Republic (see Post 116), Cameroon has no significant temperature data before 1950. However, the change in its climate is more reminiscent of that of West Africa (see Post 114). Before 1990 the climate is stable; thereafter the mean temperature appears to increase by about 0.5°C (see Fig. 117.1 below). This is a modest temperature rise and much less than the often quoted IPCC global value.

 

Fig. 117.1: The mean temperature change for Cameroon since 1940 relative to the 1951-1980 monthly averages. The best fit is applied to the monthly mean data from 1956 to 1990 and has a slight negative gradient of -0.02 ± 0.20 °C per century.

 

In order to quantify the changes to the climate of Cameroon since 1940 the temperature anomalies for the fifteen stations with the most data (i.e. over 300 months of data) were determined and averaged. This was done using the usual method as outlined in Post 47 and involved first calculating the temperature anomaly each month for each station, and then averaging those anomalies to determine the mean temperature anomaly (MTA) for the country. This MTA is shown as a time series in Fig. 117.1 above and clearly shows that temperatures declined continuously from 1940.

The process of determining the MTA in Fig. 117.1 involved first determining 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 per month within the MRT interval then its anomalies were included in the calculation of the mean temperature anomaly (MTA). The total number of stations included in the MTA in Fig. 117.1 each month is indicated in Fig. 117.2 below. The peak in the frequency between 1950 and 1990 suggests that the 1951-1980 interval was probably the most appropriate to use for the MRTs.

 

Fig. 117.2: The number of station records included each month in the mean temperature anomaly (MTA) trend for Cameroon in Fig. 117.1.

 

The locations of the sixteen stations whose data was used to determine the MTA in Fig. 117.1 are shown in the map in Fig. 117.3 below. Eight are medium stations with over 480 months of data, but only one station has over 800 months of data before 2014, and only two have any data before 1950 (see here for a full list). In addition, there are another eight stations with over 300 months of data. According to the map below the geographical spread of stations is fairly uniform, in which case the simple average of the anomalies from all stations used to construct the MTA in Fig. 117.1 should yield a fairly accurate temperature trend for the country as a whole.

 

Fig. 117.3: The (approximate) locations of the sixteen longest weather station records in Cameroon. 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 medium stations with over 480 months of data, while diamonds denote stations with more than 300 months of data.

 

The MTA in Fig. 117.1 shows the temperature change over the time period where the data is most numerous and therefore reliable. There are, however, two stations with data before 1940. These are in the two main cities of Douala and Yaoundé. The fact that in both cases the data is discontinuous with large gaps in the data between 1900 and 1940, and both stations are located in large urban areas, would suggest the data for both is not representative of the country as a whole. If we do include this earlier data we get the extended MTA shown in Fig. 117.4 below. This appears to imply an additional warming of 1.3°C occurred before 1940 when increases in carbon dioxide levels were small which also raises questions about the quality of the data. For these reasons I would tend to discount all the MTA data before 1950.

 

Fig. 117.4: The mean temperature change for Cameroon since 1880 relative to the 1951-1980 monthly averages. The best fit is applied to the monthly mean data from 1956 to 1990 and has a slight negative gradient of -0.02 ± 0.20 °C per century.

 

If we next consider the change in temperature based on Berkeley Earth (BE) adjusted data we get the MTA data in Fig. 117.5 below. This again was determined by averaging each monthly anomaly from the sixteen longest stations and suggests that the climate was fairly stable from 1940 until 1980 but then warmed by about 0.7°C thereafter.

 

Fig. 117.5: Temperature trends for Cameroon based on Berkeley Earth adjusted data. The best fit linear trend line (in red) is for the period 1952-2011 and has a positive gradient of +1.46 ± 0.05°C/century.

 

Comparing the curves in Fig. 117.5 with the published Berkeley Earth (BE) version for Cameroon in Fig. 117.6 below shows that there is good agreement between the two sets of data. This indicates that the simple averaging of anomalies used to generate the BE MTA in Fig. 117.5 is as effective and accurate as the more complex gridding method used by Berkeley Earth in Fig. 117.6. In which case simple averaging should be just as effective and accurate in generating the MTA using raw unadjusted data in Fig. 117.1. How Berkeley Earth managed to determine the temperature change in Cameroon between 1900 and 1950 in Fig. 17.6 when there is virtually no data for this period is a point of debate.

 

Fig. 117.6: The temperature trend for Cameroon since 1840 according to Berkeley Earth.

 

The differences between the MTA in Fig. 117.4 and the BE versions using adjusted data in Fig. 117.5  are instead 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. 117.4 and Fig. 117.5. The magnitudes of these adjustments are shown graphically in Fig. 117.7 below. The blue curve is the difference in MTA values between adjusted (Fig. 117.5) and unadjusted data (Fig. 117.4), while the orange curve is the contribution to those adjustments arising solely from breakpoint adjustments. Neither are larger than about 0.2°C.

 

Fig. 117.7: The contribution of Berkeley Earth (BE) adjustments to the anomaly data in Fig. 117.5 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 1952-2011 has a positive gradient of +0.41 ± 0.03 °C per century. The orange curve shows the contribution just from breakpoint adjustments.

 

Summary

According to the raw unadjusted temperature data, the climate of Cameroon was stable until 1990 and then warmed by between 0.3°C and 0.6°C (see Fig. 117.1).

Over the same period adjusted temperature data from Berkeley Earth appears to show that the climate of Cameroon has warmed by about 0.8°C (see Fig. 117.5).


Acronyms

BE = Berkeley Earth.

MRT = monthly reference temperature (see Post 47).

MTA = mean temperature anomaly.

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

116: Central African Republic (CAR) - temperature trends COOLING

There are thirteen medium stations with over 480 months of data in the Central African Republic (CAR), but only one station has over 800 months of data before 2014, and none have any data before 1940 (see here for a full list). In addition, there are another two stations with over 300 months of data.

The neighbouring countries of Chad, Cameroon, Congo and the DRC (formerly Zaire) have virtually no data before 1940 either. Only Sudan has significant data pre-1940. This means it is not possible to know the true temperature trend of CAR before 1940. What the data that we do have tells us is that the climate of the Central African Republic cooled by over 0.5°C from 1950 onwards (see Fig. 116.1 below).

 

Fig. 116.1: The mean temperature change for the Central African Republic relative to the 1961-1990 monthly averages. The best fit is applied to the monthly mean data from 1941 to 2005 and has a negative gradient of -0.50 ± 0.11 °C per century.

 

In order to quantify the changes to the climate of the CAR since 1940 the temperature anomalies for each of the fifteen stations with the most data were determined and averaged. This was done using the usual method as outlined in Post 47 and involved first calculating the temperature anomaly each month for each station, and then averaging those anomalies to determine the mean temperature anomaly (MTA) for the country. This MTA is shown as a time series in Fig. 116.1 above and clearly shows that temperatures declined continuously from 1940.

The process of determining the MTA in Fig. 116.1 involved first determining the monthly reference temperatures (MRTs) for each station using a set reference period, in this case from 1961 to 1990, and then subtracting the MRTs from the raw temperature data to deliver the anomalies. If a station had at least twelve valid temperatures per month within the MRT interval then its anomalies were included in the calculation of the mean temperature anomaly (MTA). The total number of stations included in the MTA in Fig. 116.1 each month is indicated in Fig. 116.2 below. The peak in the frequency between 1960 and 1990 suggests that the 1961-1990 interval was indeed the most appropriate to use for the MRTs.

 

Fig. 116.2: The number of station records included each month in the mean temperature anomaly (MTA) trend for the Central African Republic in Fig. 116.1.

 

The locations of the fifteen stations with the most temperature data are shown in the map in Fig. 116.3 below. This appears to show that the geographical spread is fairly uniform, although there does appear to be more stations in the south of the country than in the north. This variation in station density is probably not sufficient to significantly distort the average in Fig. 116.1 from its true value though. In which case the simple average of the anomalies from all stations used to construct the MTA in Fig. 116.1 should still yield a fairly accurate temperature trend for the country as a whole.

 

Fig. 116.3: The (approximate) locations of the fifteen longest weather station records in the Central African Republic. 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 medium stations with over 480 months of data, while diamonds denote stations with more than 250 months of data.

 

If we next consider the change in temperature based on Berkeley Earth (BE) adjusted data we get the MTA data in Fig. 116.4 below. This again was determined by averaging each monthly anomaly from the fifteen longest stations and suggests that the climate was fairly stable before 1980 but then warmed by about 0.75°C thereafter.

 

Fig. 116.4: Temperature trends for the Central African Republic based on Berkeley Earth adjusted data. The best fit linear trend line (in red) is for the period 1941-2010 and has a positive gradient of +0.83 ± 0.08°C/century.

 

If we next compare the curves in Fig. 116.4 with the published Berkeley Earth (BE) version for the CAR in Fig. 116.5 below we see that there is good agreement between the two sets of data after 1940. This indicates that the simple averaging of anomalies used to generate the BE MTA in Fig. 116.4 is as effective and accurate as the more complex gridding method used by Berkeley Earth in Fig. 116.5. In which case simple averaging should be just as effective and accurate in generating the MTA using raw unadjusted data in Fig. 116.1. What is more difficult to explain is how Berkeley Earth have determined the climate for the CAR as far back as 1880 when there is virtually no reliable temperature data for the country before 1940.

 

Fig. 116.5: The temperature trend for the Central African Republic since 1850 according to Berkeley Earth.

 

The differences between the MTA in Fig. 116.1 and the BE version using adjusted data in Fig. 116.4  are probably 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. 116.1 and Fig. 116.4. The magnitudes of these adjustments are shown graphically in Fig. 116.6 below. The blue curve is the difference in MTA values between adjusted (Fig. 116.4) and unadjusted data (Fig. 116.1), 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 1940 of up to 1.5°C.

 

Fig. 116.6: The contribution of Berkeley Earth (BE) adjustments to the anomaly data in Fig. 116.4 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 1941-2010 has a positive gradient of +1.32 ± 0.03 °C per century. The orange curve shows the contribution just from breakpoint adjustments.

Summary

According to the raw unadjusted temperature data, the climate of the CAR has cooled from 1940 by about 0.5°C (see Fig. 116.1).

Over the same period adjusted temperature data from Berkeley Earth appears to show that the climate of the CAR has warmed by over 0.5°C (see Fig. 116.5).

 


Acronyms

BE = Berkeley Earth.

MRT = monthly reference temperature (see Post 47).

MTA = mean temperature anomaly.

Link to list of all stations in the Central African Republic and their raw data files.

115: Chad - temperature trends WARMING 2°C after 1980

Any analysis of the climate of Chad is complicated by two factors: the first is the lack of data; the second is the gap in the temperature data from 1979 to 1986 that coincides with a sudden jump in temperatures. The longest temperature record for Chad has less than 800 months of data before 2014 and only extends back to 1941. In total there are only seven medium stations with over 480 months of data and a further five stations with over 300 months of data. The locations of these stations are shown on the map in Fig. 115.1 below. All but one of these twelve stations are in the southern half of the country.

Fig. 115.1: The (approximate) locations of the twelve longest weather station records in Chad. 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 medium stations with over 480 months of data, while diamonds denote stations with more than 300 months of data.

In order to quantify the changes to the climate of Chad since 1941 the temperature anomalies for each of the twelve stations shown in Fig. 115.1 were determined and averaged. This was done using the usual method as outlined in Post 47 and involved first calculating the temperature anomaly each month for each station, and then averaging those anomalies to determine the mean temperature anomaly (MTA) for the country. This MTA is shown as a time series in Fig. 115.2 below and clearly shows that temperatures declined slowly before 1980 by about 0.4°C in total, and rose more rapidly by up to 2°C thereafter. However, the large gap in data from 1979 to 1987, together with the abrupt jump in temperatures after 1987, together raise questions over the reliability of the post-1987 data.

Fig. 115.2: The mean temperature change for Chad relative to the 1951-1980 monthly averages. The best fit is applied to the monthly mean data from 1941 to 1975 and has a negative gradient of -1.04 ± 0.38 °C per century.

The process of determining the MTA in Fig. 115.2 involved first determining 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 per month within the MRT interval then its anomalies were included in the calculation of the mean temperature anomaly (MTA). The total number of stations included in the MTA in Fig. 115.2 each month is indicated in Fig. 115.3 below. The peak in the frequency between 1950 and 1980 suggests that the 1951-1980 interval was indeed the most appropriate to use for the MRTs.

Fig. 115.3: The number of station records included each month in the mean temperature anomaly (MTA) trend for Chad in Fig. 115.2.

If we next consider the change in temperature based on Berkeley Earth (BE) adjusted data we get the MTA data in Fig. 115.4 below. This again was determined by averaging each monthly anomaly from all the available stations and suggests that the climate was stable before 1980 and only warmed by about 1°C thereafter. So there is less cooling before 1980 and less warming after. It also appears that the data after 1987 has been offset downwards by about 0.5°C in order to eliminate some of the temperature rise seen in Fig. 115.2.

Fig. 115.4: Temperature trends for Chad based on Berkeley Earth adjusted data. The best fit linear trend line (in red) is for the period 1941-1975 and has a very slight negative gradient of -0.04 ± 0.16°C/century.

If we compare the curves in Fig. 115.4 with the published Berkeley Earth (BE) version in Fig. 115.5 below we see that there is good agreement between the two sets of data at least as far back as 1950. This indicates that the simple averaging of anomalies used to generate the BE MTA in Fig. 115.4 gives similar results to the more complex gridding method used by Berkeley Earth in Fig. 115.5. The official BE temperature trend also claims to know the temperature trend before 1940 and as far back as 1870, and also for all of the 1980s, even though there is no actual data in Chad for either time period.

Fig. 115.5: The temperature trend for Chad since 1850 according to Berkeley Earth.

The differences between the MTA in Fig. 115.2 and the BE version using adjusted data in Fig. 115.4  are instead 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. 115.2 and Fig. 115.4. The magnitudes of these adjustments are shown graphically in Fig. 115.6 below. The blue curve is the difference in MTA values between adjusted (Fig. 115.4) and unadjusted data (Fig. 115.2), while the orange curve is the contribution to those adjustments arising solely from breakpoint adjustments. It can be seen that the main adjustment is a vertical offset of data after 1987.

Fig. 115.6: The contribution of Berkeley Earth (BE) adjustments to the anomaly data in Fig. 115.4 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 has a negative gradient of -1.17 ± 0.04 °C per century. The orange curve shows the contribution just from breakpoint adjustments.

Summary

According to the raw unadjusted temperature data, the climate of Chad cooled by 0.4°C until 1975 and then warmed by about 2°C (see Fig. 115.2). However, the gap in data from 1979-1987 and the subsequent jump in temperatures thereafter may mean the rise post-1987 is only about 1.4°C and the net rise only 1°C.

Over the same period adjusted temperature data from Berkeley Earth appears to show that the climate of Chad has warmed by about 1.0°C (see Fig. 115.4). This is possibly the same as the raw data overall, but the pattern over time is different.


Acronyms

BE = Berkeley Earth.

MRT = monthly reference temperature (see Post 47).

MTA = mean temperature anomaly.

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

114: West Africa - temperature trends WARMING 0.7°C after 1980

There are fifteen countries that together make up West Africa: Mauritania, Mali, Burkina Faso, Niger, Nigeria, Benin, Togo, Ghana, Ivory Coast (Côte d'Ivoire), Liberia, Sierra Leone, Guinea, Guinea-Bissau, Senegal, and The Gambia. Unfortunately there are only seven long stations in the entire region with over 1200 months of data, most of which have cooling trends (see Fig. 114.1 below). There are, however, another 111 medium stations with over 480 months of data. Of these nearly 100 have over 600 months of data, nearly 50 have over 800 months of data and 11 have over 1000 months of data. For a full list see here.

While the region as a whole has reasonably good data, most individual countries do not. Most countries have fewer than ten long and medium stations in total. For that reason it makes sense to determine the temperature change across the region as a whole rather than concentrating on the data from individual countries. On the positive side, the distribution of stations is fairly uniform so averaging of the mean temperature anomaly (MTA) data should yield accurate results for the temperature change at least for the most recent 800 months (i.e. as far back as 1950), and possibly as far back as 1920.

Fig. 114.1: The (approximate) locations of the 118 longest weather station records in West Africa. 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.

The temperature anomalies for each station in Fig. 114.1 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 per month within the MRT interval then its anomalies were included in the calculation of the regional mean temperature anomaly (MTA). The resulting MTA after 1900 is shown in Fig. 114.2 below. It can be seen that before 1975 the climate was slowly cooling with the mean temperature falling by about 0.2°C over the preceding 75 years. Then after 1975 there was a rapid warming of about 0.75°C. The net warming since 1900 is therefore just over 0.5°C.

Fig. 114.2: The mean temperature change for West Africa since 1880 relative to the 1951-1980 monthly averages. The best fit is applied to the monthly mean data from 1901 to 1970 and has a negative gradient of -0.25 ± 0.08 °C per century.

The total number of stations included in the MTA in Fig. 114.2 each month is indicated in Fig. 114.3 below. The peak in the frequency around 1970 suggests that the 1951-1980 interval for the MRTs was indeed the most appropriate to use. It also indicates that there are a couple of stations with data before 1880 and about six with data before 1900.

Fig. 114.3: The number of station records included each month in the mean temperature anomaly (MTA) trend for West Africa in Fig. 114.2.

 

Unfortunately the data before 1880 is fragmented with large fluctuations in its values, as the MTA in Fig. 114.4 below shows. It is therefore likely to be very unreliable and thus adds nothing to our understanding of climate change in the region.

Fig. 114.4: The mean temperature change for West Africa since 1840. The best fit has a slight negative gradient of -0.25 ± 0.08 °C per century.

In contrast to Fig. 114.2, the corresponding MTA dataset based on data that has been adjusted by Berkeley Earth (BE) exhibits a strong warming trend before 1975 with temperatures rising by over 1.0°C since 1900 (see Fig. 114.5 below).

Fig. 114.5: Temperature trends for West Africa based on Berkeley Earth adjusted data. The best fit linear trend line (in red) is for the period 1901-2010 and has a gradient of +0.85 ± 0.03°C/century.

If we next compare the curves in Fig. 114.5 with the published Berkeley Earth (BE) version for West Africa in Fig. 114.6 below we see that there is remarkably good agreement between the two sets of data at least as far back as 1900. This indicates that the simple averaging of anomalies used to generate the BE MTA in Fig. 114.5 is as effective and accurate as the more complex gridding method used by Berkeley Earth in Fig. 114.6. In which case simple averaging should be just as effective and accurate in generating the MTA using raw unadjusted data in Fig. 114.2. This also suggests that any difference between the two averages cannot be due primarily to the averaging process, but must instead be due at least in part to the temperature adjustments made by Berkeley Earth.

Fig. 114.6: The temperature trend for West Africa since 1840 according to Berkeley Earth.

The differences between the MTA in Fig. 114.2 and the BE versions using adjusted data in Fig. 114.5  are instead 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. 114.2 and Fig. 114.5. The magnitudes of these adjustments are shown graphically in Fig. 114.7 below.  

Fig. 114.7: The contribution of Berkeley Earth (BE) adjustments to the anomaly data in Fig. 114.5 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 1921-2010 has a slight negative gradient of -0.057 ± 0.009 °C per century. The orange curve shows the contribution just from breakpoint adjustments.

The blue curve in Fig. 114.7 above is the difference in MTA values between adjusted (Fig. 114.5) and unadjusted data (Fig. 114.2), while the orange curve is the contribution to those adjustments arising solely from breakpoint adjustments. Both adjustments are negligible after 1920, but before 1920 they add an additional warming of up to 0.8°C. But as Fig. 114.3 shows, the MTA data before 1920 is based on temperature data from at most 21 stations. So how reliable is the raw data before 1920? And how reliable are the BE adjustments before 1920?

The case in favour of believing the raw unadjusted data before 1920 is two-fold: that it is the real data, and that it follows the same trend as the raw data after 1920. In addition, the case against the adjusted data is that the adjustments are so large relative to any made after 1920. In reality, it is difficult to say categorically which is the more reliable, but the unadjusted data does also correlate with raw unadjusted data from elsewhere, such as southern Africa (see Posts 37, 77, 78, 79) and the Southern Hemisphere (see Post 64).

Summary

According to the raw unadjusted temperature data, the climate of West Africa cooled until 1975 and then warmed by about 0.75°C (see Fig. 114.2). The net warming since 1900 is only about 0.5°C.

Over the same period adjusted temperature data from Berkeley Earth appears to show that the climate of West Africa has warmed by over 1.0°C (see Fig. 114.5).

The data before 1920 is based on a small sample (only 21 stations) and so could be considered highly uncertain.


Acronyms

BE = Berkeley Earth.

MRT = monthly reference temperature (see Post 47).

MTA = mean temperature anomaly.

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

100. List of completed temperature analyses by country

 

As this is my 100th post on this blog I thought it would be a good moment to summarize the results that have emerged from the temperature data I have analysed so far. Below is a list of all the countries and regions that I have investigated to date with links to the relevant post. This amounts to about 60 countries, states and territories in total, which is roughly one third of all the countries in the world. 

The main areas that so far remain to be studied are the Arctic, Canada, Russia, UK, Scandinavia, Mediterranean, North Africa, Middle East, China and Japan. In the Southern Hemisphere only Brazil, Venezuela, Guyana, Suriname, French Guiana and the South Atlantic remain. However, most of these Southern Hemisphere regions are already included in the analysis of South America in Post 35.

Europe

Europe has the longest temperature records available with several in Germany, Sweden and the Netherlands stretching back to the early 18th century. The average of the 109 longest records yields a mean temperature anomaly (MTA) that shows a small but continuous warming of about 0.1°C per century for over 200 years until 1988. Then the temperature jumps suddenly by over 1°C. The reason for this jump is unclear. It is certainly not related directly to carbon dioxide emissions. The only countries that appear to have strong warming trends are the Benelux countries, Denmark and Switzerland. The Baltic states and most of central Europe appear to cool before 1980 and then warm suddenly.

109 longest station records (Post 44)

Austria (Post 55)

Baltic States (Post 51)

Belgium and Luxembourg (Post 40)

Central Europe average (Post 57)

Czechoslovakia (Post 53)

Denmark (Post 48)

Germany (Post 49)

Hungary (Post 54)

Netherlands (Post 41)

Poland (Post 50)

Switzerland (Post 56)

 

USA

The USA may not have any temperature records that are as long as the longest that Europe can boast, but its temperature data from 1850 onwards is the best there is. Virtually every state has over 100 station records with over 50 years of data and over 50 records with over 100 years of data. An average of the 400 longest temperature records appears to indicate that the climate warmed by more than 2°C from 1780 to 1920 when carbon dioxide levels barely increased, and then cooled by over 0.5°C when carbon dioxide levels took off. The early warming cannot therefore be due to CO2 and is therefore generally attributed to urbanization and deforestation in the north and east. The cooling seen after 1920 is also seen in most southern states like Louisiana, Mississippi and Texas.

400 longest station records (Post 66)

Louisiana (Post 97)

Mississippi (Post 99)

Texas (Post 52)

 

Central America

The temperature data for Central America can basically be split between Mexico and the rest, however, even then there are more than four times as many stations in Mexico as there are in the rest of Central America. The picture in Mexico is also complicated by the stations there falling into two distinct types from two different sources. On balance it is likely that the overall climate was stable until 1980 and then warmed over the following twenty years by about 1°C.

Mexico (Post 93)

Rest of Central America (Post 94)

South America

Of the countries in South America studied so far, only Argentina (0.6°C), Ecuador (1°C) and Uruguay (1°C) show significant warming, although the Ecuador data is far from reliable. In Paraguay and Chile the climate has cooled while in most other countries it has remained stable. The average of all medium and long stations in South America yields a warming of about 0.5°C since 1900.

All long and medium stations (Post 35)

Argentina (Post 61)

Bolivia (Post 58)

Chile (Post 62)

Colombia (Post 95)

Ecuador (Post 96)

Paraguay (Post 59)

Peru (Post 63)

Uruguay (Post 60)

Asia

My analysis so far of temperatures in Asia has focused on the countries of Indochina and the Indian subcontinent. The overall temperature trend for Indochina is one of cooling before 1980 and warming thereafter. The result is that temperatures in 2010 are barely any higher than they were in 1890. This is also reflected in the individual temperature records of Burma, Malaysia and Vietnam, while those of the Philippines and Thailand remain stable from 1920 onwards. In India and Pakistan there is little warming before 1990 and then a sudden jump in temperatures of about 0.5°C in the mid-1990s. For Sri Lanka the jump in temperature occurs in 1978 while Bangladesh sees a continuous warming of only 0.3°C per century. 

Bangladesh (Post 74)

Burma/Myanmar (Post 69)

India including Nepal (Post 71)

Indian subcontinent (Post 75)

Indochina (Post 70)

Malaysia and Singapore (Post 69)

Pakistan (Post 72)

Philippines (Post 69)

Sri Lanka (Post 73)

Thailand (Post 69)

Vietnam (Post 69)

Africa

Most of southern Africa has exhibited some significant warming of over 1°C since 1980 but the overall picture before 1980 is varied. Angola, Mozambique and South Africa show no warming before 1980 while Malawi, Zambia, Zimbabwe and Madagascar all cool significantly by as much as they later warm. The data for Namibia and Botswana is not great but may indicate a slight warming before 1980 as well as much larger warming thereafter. Of all the countries listed below, Madagascar, Mozambique, South Africa and Zimbabwe have the best quality data and none of these countries appear to exhibit any warming before 1980.

Angola (Post 82)

Botswana (Post 38)

Madagascar (Post 77)

Mozambique (Post 78)

Namibia (Post 39)

South Africa including Lesotho and Eswatini/Swaziland (Post 37)

Zambia and Malawi (Post 81)

Zimbabwe (Post 79)

Australia

Analysis of temperature data for Australia indicates that the mean temperature trend is parabolic with the climate cooling from 1875 to 1960 and then warming. Overall temperatures in 2010 are only about 0.1-0.2°C warmer than in 1875 with temperatures having increased by about 0.5°C since 1960. This pattern in seen in most states such as South Australia, New South Wales and Victoria. It is harder to be conclusive for Tasmania and Western Australia due to a lack of data before 1900 while the trend in Northern Territory is one of consistent cooling. Only Queensland shows constant warming of about 1°C since 1990.

Australia (Post 26)

New South Wales and ACT (Post 18)

Northern Territory (Post 23)

Queensland (Post 24)

South Australia (Post 21)

Tasmania (Post 20)

Victoria (Post 19)

Western Australia (Post 22)

Oceania

Most of the countries and regions of Oceania show little of no warming. In Antarctica the only warming is found around the peninsula. New Zealand cools slightly from 1860 until 1960 then warms by about 0.5°C, rather like much of Australia. Yet despite this, temperatures in 2010 are barely above those in 1860. In Indonesia only the capital city Jakarta shows any strong warming but the average temperature for the country remains stable, although data quality and quantity before 1960 is poor. This is also true for Papua New Guinea where there is some evidence of warming after 1960 by about 0.5°C. In the South Pacific there is a contrast between east and west with the eastern half cooling significantly while the west cools slightly until 1970 before warming again by about 0.5°C. In fact of all the regions listed below, only the Indian Ocean shows significant warming of about 1°C.

Antarctica (Post 30)

Indian Ocean (Post 76)

Indonesia (Post 31)

New Zealand (Post 8)

Papua New Guinea (Post 32)

South Pacific Islands - East (Post 34)

South Pacific Islands - West (Post 33)

Southern Hemisphere

An average of the temperature anomalies from the 1000 longest records in the Southern Hemisphere shows a slight cooling of about 0.1°C until 1975 followed by a modest warming of only about 0.6°C.

Southern Hemisphere station average (Post 64)

84. Southern Africa - a summary of temperature trends

Over the last fifteen months I have analysed the temperature data for nine different countries in southern Africa and calculated their temperature trends. My analysis began in September 2020 with South Africa (Post 37) before continuing on to Botswana (Post 38) and Namibia (Post 39) in October. More recently I have examined the temperature records of Madagascar (Post 77), Mozambique (Post 78), Zimbabwe (Post 79), Zambia and Malawi (Post 81), and Angola (Post 82).

The one constant for all these countries is the temperature trend since 1980. In almost all cases the climate shows evidence of warming, with the temperature rise varying between 0.5°C and 1.5°C. Before 1980 the picture is different with little or no temperature rise being seen, and in many cases (e.g. Madagascar, Zimbabwe, Zambia and Malawi) a significant cooling taking place. These results are summarized in the graphs below.

The main conclusions to be drawn here are as follows.

• All the warming in southern Africa has occurred since 1980.
• There is probably no warming before 1980, and possibly some cooling.
• The net warming since 1850 is likely to be between 0.5°C and 1.0°C. This is less than the IPCC global value of 1.3°C, and much less than the land-based Berkeley Earth average of almost 2.0°C (see Fig. 80.1 in Post 80).

In the New Year I will return to the African data and analyse the data from countries in sub-Saharan and equitorial countries. In the meantime I will return to considering the physics of the Greenhouse Effect that I introduced in Post 12 (black-body radiation) and Post 13 (Earth's energy budget). In particular I will look to answer two questions: why is Venus so hot, and why is Mars so cold?

Fig. 37.2: The mean temperature anomaly (MTA) for South Africa since 1840. The best fit is applied to the interval 1857-1976 and has a gradient of +0.017 ± 0.056 °C per century. The monthly temperature changes are relative to the 1961-1990 monthly averages.

Fig. 38.3: The mean temperature anomaly (MTA) for Botswana since 1917. The best fit is applied to the interval 1917-1976 and has a positive gradient of +0.72 ± 0.26 °C per century. The monthly temperature changes are defined relative to the 1961-1990 monthly averages.

Fig. 39.8: The mean temperature anomaly (MTA) for Namibia since 1885. The best fit is applied to the interval 1944-2001 and has a positive gradient of +2.28 ± 0.17 °C per century. The monthly temperature changes are defined relative to the 1971-1990 monthly averages.

Fig. 77.6: The mean temperature anomaly (MTA) for Madagascar. The best fit is applied to the monthly mean data from 1932 to 2011 and has a negative gradient of -0.15 ± 0.07 °C per century.

Fig. 78.6: The mean temperature anomaly (MTA) relative to the 1931-1960 monthly averages for stations in Mozambique with over 300 months of data but excluding Lourenço Marques. The best fit is applied to the monthly mean data from 1921 to 1980 and has a positive gradient of +0.16 ± 0.12 °C per century.

Fig. 79.2: The mean temperature anomaly (MTA) for Zimbabwe relative to the 1971-2000 monthly averages based on an average of anomalies from stations with over 360 months of data. The best fit is applied to the monthly mean data from 1916 to 1975 and has a negative gradient of -0.96 ± 0.20 °C per century.

Fig. 81.2: The mean temperature anomaly (MTA) relative to the 1951-1980 monthly averages based on an average of anomalies from stations with over 360 months of data. The best fit is applied to the monthly mean data from 1921 to 1975 and has a negative gradient of -2.72 ± 0.17 °C per century.

Fig. 82.5: The mean temperature anomaly (MTA) for Angola relative to the 1951-1980 monthly averages based on an average of anomalies from stations with over 300 months of data. The best fit is applied to the monthly mean data from 1941 to 1980 and has a positive gradient of 0.18 ± 0.17 °C per century.

82. Angola - temperature trends STABLE

The biggest problem we are confronted with when it comes to determining the extent of climate change in Angola is a lack of temperature data. There is only one long station with over 1200 months of data and another four medium stations with over 480 months of data. Only two stations have a significant quantity of data before 1939, and they are both in the capital Luanda; and none have any reliable data after 1980 due to the protracted civil war. A full list of stations for Angola can be found here.

If we include stations with over 240 months of data, then there are a total of eighteen temperature records that we can use. Their locations are indicated on the map below in Fig. 82.1. Overall they are fairly evenly distributed across the country, although the stations with the longest temperature records are generally found nearest the coast. The consequence of this deficit of data is that it is impossible to definitively determine the temperature trend for the country either before 1940, or after 1980. Between 1940 and 1980 the data suggests that no warming took place.

Fig. 82.1: The (approximate) locations of the main weather stations in Angola. Those stations with a high warming trend between 1901 and 2000 are marked in red while those with a cooling or stable trend are marked in blue.

The monthly anomalies for each station were created in the usual manner, as outlined in Post 47. First a suitable thirty year interval was chosen for calculating the monthly reference temperatures (MRTs). In this case the period 1951-1980 was chosen as that corresponded to the interval that overlapped with the maximum number of station records. The twelve MRTs for each station dataset were calculated for each of the twelve months by averaging the monthly temperatures in the reference period for that station. The MRTs were then subtracted from all the respective monthly temperature data for that station to generate the anomalies. The anomalies from all the stations were then averaged to give the mean temperature anomaly (MTA) for the region in that month. Employing a simple average of the station data rather than using Kriging, homogenization and gridding is sufficiently accurate if the stations are fair evenly distributed, which the map in Fig. 82.1 suggests to be the case. The resulting mean temperature anomaly since 1875 is shown below in Fig. 82.2.

Fig. 82.2: The mean temperature anomaly (MTA) for Angola relative to the 1951-1980 monthly averages based on an average of anomalies from stations with over 300 months of data. The best fit is applied to the monthly mean data from 1881 to 1980 and has a positive gradient of 1.39 ± 0.08 °C per century.

The MTA in Fig. 82.2 exhibits a similar warming profile to the IPCC global trends, but this is somewhat misleading. The data before 1939 is based on only two stations, and both of these are situated in the capital Luanda (Berkeley Earth ID 151475 and 2824), while the data after 1980 is highly sporadic and discontinuous for all stations with data in this period. This is illustrated in Fig. 82.3 below which shows the station frequency count in the MTA in Fig. 82.2. 

Fig. 82.3: The number of station records included each month in the mean temperature anomaly (MTA) trend in Fig. 82.2.

The station with the longest data set is Luanda (Berkeley Earth ID 151475), the monthly anomalies of which are shown in Fig. 82.4 below. The second station in Luanda (Berkeley Earth ID 2824) has no data after 1973, while its data after 1951 is virtually identical to that of station 151475. Before 1951, however, the data from the two stations disagree by an average of almost 0.5°C. This hints that the two sets of station data may not be completely independent, or may be the result of combining other datasets.

Fig. 82.4: The temperature anomaly for Luanda relative to the 1951-1980 monthly averages. The best fit is applied to all the monthly mean data and has a positive gradient of 1.63 ± 0.06 °C per century.

The main conclusion to be drawn from Fig. 82.3 is that both the data before 1939 and the data after 1980 are unreliable. The data after 1980 is unreliable because there is so little of it, and it is highly fragmented. The data before 1939 is unreliable because it is based on just two stations that are located in the most highly populated region of the country. While these two stations do corroborate each other, particularly after 1951, they are still unlikely to be representative of the climate of the country as a whole. We know this because we see it in many countries and regions: strong warming in the major cities like Jakarta (see Post 31), Sydney (BE-151986) in NSW (Post 18) and Melbourne (BE-151813) in Victoria (Post 19), but a totally different temperature trend in the surrounding region.

Fig. 82.5: The mean temperature anomaly (MTA) for Angola relative to the 1951-1980 monthly averages based on an average of anomalies from stations with over 300 months of data. The best fit is applied to the monthly mean data from 1941 to 1980 and has a positive gradient of 0.18 ± 0.17 °C per century.

If we therefore just look at the data from 1940 to 1980 we see that there is no significant warming (see Fig. 82.5 above). The slight rise in the best fit line is at least five times less than the standard deviation of the MTA, and is almost the same as the uncertainty in the best fit. And this is also similar to the extent of the warming claimed by Berkeley Earth for this period (see Fig. 82.6 below).

Fig. 82.6: The temperature trend for Angola since 1840 according to Berkeley Earth.

However, the temperature trend for Angola according to Berkeley Earth (BE) does exhibit strong warming both before 1939 and after 1980. While the latter is certainly a reasonable assumption based on trends elsewhere in the region, it is not a conclusion that can be confidently presented based on the available data. As for the data in Fig. 82.6 before 1939, this may be supported by the station data from Luanda, but it differs significantly from the trends I have determined for neighbouring countries such as Namibia (Post 39), Botswana (Post 38), Zambia (Post 81) and Zimbabwe (Post 79).

Summary

Temperatures in Angola between 1939 and 1980 appear to be stable.

There is insufficient data to determine the extent of climate change in Angola either before 1939, or after 1980.

Acronyms

BE = Berkeley Earth.

MRT = monthly reference temperature (see Post 47).

MTA = mean temperature anomaly.