These graphs show average historic annual temperature and precipitation for the village of At-Bashy. This data comes from stations maintained as part of a much larger network of weather observation stations across the former USSR. Precipitation data dates back to 1927 and temperature data for At-Bashy dates back to 1962, however both datasets have incomplete years of data that were excluded from my graphs. The data does not continue to present day, as most weather stations were abandoned in 1990. In fact today, there are less than 25 weather stations in Kyrgyzstan, just a fraction of what existed several decades ago.
According to the regressions, which I have not analyzed their statistical significance, the data shows an increasing temperature trend and a decreasing precipitation trend. The significance of these trends should not be extended to make generalization about regional trends over the past century. However the data still provides an insight and when coupled with additional data sites will provide a broader glimpse of trends.
Obtaining the data was a process in itself. During the winter I contacted the Kyrgyz Hydro- meteorological Center certain they would be thrilled to provide me with the data I was interested. To my surprise I was informed that the data would be available at $10 US Dollar per year per station. That is a hefty price tag of $4200 for just 4 stations’ data! Unwilling to pay, as I made more contacts I thankfully obtained the data free of charge from several sources including Dr. Ryskul Usubaliev at Central Asia Institute of Applied Geosciences and also from Dr. Chiyuki Narama at Nagoya University in Japan. The problem was there were discrepancies and the data did not match. Recently, Dr. Vladimir Aizen at the University of Idaho made a majority of historic USSR meteorological data available via a web data server at http://www.sci.uidaho.edu/cae/data/index.html. Dr. Aizen and Dr. Narama have both assured me to the accuracy of their data, and being that there data is identical, it is the data I have chosen to use.
The uncertainty in data accuracy is significant because each data value can have a huge effect on the resulting trend. One small difference can greatly exaggerate an increasing trend that could be used to falsely indicate warming. On the reverse, another error could negate an actual positive trend.
This story highlights the delicate nature of numbers and data, as well as questioning the significance that is placed on the results. Beyond questioning the data sources and transcription, I have to wonder about the accuracy of the observation records. Although error in transcription from original sources to digital sources can be overcome, there is no way to overcome original biases from measurement or recorder error.
I will still use the data in my research, but it will be supportive documentation. Coupling the trends of multiple stations in the area should provide a better representation. Additionally, I will be interested to see any correlation between the data and locals’ observations.
Obtaining the data was a process in itself. During the winter I contacted the Kyrgyz Hydro- meteorological Center certain they would be thrilled to provide me with the data I was interested. To my surprise I was informed that the data would be available at $10 US Dollar per year per station. That is a hefty price tag of $4200 for just 4 stations’ data! Unwilling to pay, as I made more contacts I thankfully obtained the data free of charge from several sources including Dr. Ryskul Usubaliev at Central Asia Institute of Applied Geosciences and also from Dr. Chiyuki Narama at Nagoya University in Japan. The problem was there were discrepancies and the data did not match. Recently, Dr. Vladimir Aizen at the University of Idaho made a majority of historic USSR meteorological data available via a web data server at http://www.sci.uidaho.edu/cae/data/index.html. Dr. Aizen and Dr. Narama have both assured me to the accuracy of their data, and being that there data is identical, it is the data I have chosen to use.
The uncertainty in data accuracy is significant because each data value can have a huge effect on the resulting trend. One small difference can greatly exaggerate an increasing trend that could be used to falsely indicate warming. On the reverse, another error could negate an actual positive trend.
This story highlights the delicate nature of numbers and data, as well as questioning the significance that is placed on the results. Beyond questioning the data sources and transcription, I have to wonder about the accuracy of the observation records. Although error in transcription from original sources to digital sources can be overcome, there is no way to overcome original biases from measurement or recorder error.
I will still use the data in my research, but it will be supportive documentation. Coupling the trends of multiple stations in the area should provide a better representation. Additionally, I will be interested to see any correlation between the data and locals’ observations.
No comments:
Post a Comment