I had a look at a number of homogeneous stations (stations with no major problems) to retrieve some statistical properties. I used the monthly means of the daily maximum temperature for about 30 stations, covering as much as possible 1920-2009. I examined the monthly departures from their long-term means (12 monthly means) so each series had about 1080 data points. Overall, the best-fit linear trend ranged from 0.3 to 1.5°C for 1920-2009. Once the trend was removed, the standard deviation ranged from 2.5 to 3.5 and the autocorrelation from 0.15 to 0.25. The reverse process can be used to generate datasets. However, this approach does not capture the long-term variation, such as warming from the 1900s to the 1930s, followed by a cooling to the 1970s, and warming to the 2000s.
The types of “inhomogeneities” that I have seen in temperature series are:
1. change in annual means: 2 or 3 steps are common over 90 years; a more complex situation would be a step every 10 years
2. change in monthly means: here, the magnitude of the step would be different depending of the month (e.g. an instrument relocation could generate a positive step in the summer and a negative step in the winter)
3. change in mean and variance due to a change of instrument or observing practices
4. change in mean followed by a change in trend direction: this occurs rarely and I can’t think of a physical process that would generate this situation in temperature series.
These “inhomogeneities” could be detected using a network of neighbour stations which could be highly (or not) correlated with the tested series. In addition, the same (or other) “inhomogeneities” could be found in the neighbouring series.
3 comments:
Statistics on inhomogeneities sizes and frequencies
Hi chaps, Sorry for not being so active so far.
Concerning size and frequencies of inhomogeneities, you may check Hannart and Naveau article, where you have distributions for both, computed on french data.
Bayesian multiple change points and segmentation:
Application to homogenization of climatic series
A. Hannart and P. Naveau
WATER RESOURCES RESEARCH, VOL. 45, W10444, doi:10.1029/2008WR007689, 2009 Olivier
The monthly change being positive in summer/negative in winter is interesting and very difficult to detect. Can you give any more info on the common physical causes of any common changes? In fact trying to tie physical mechanisms to breakpoint characteristics would be really useful. I've had a quick go below and will add more when I get a moment to dig through some papers:
station move from part-exposed location (North facing wall) to exposed location (WMO standard site) = could be warmer in summer as the sensors are no longer in the shade of the north facing wall or close to its mass which will likely be cooler than the air temperature, could be cooler in winter where it is now more exposed or would the shade/cold north facing wall effect dominate?
changes in shelter type and thermometer type have been documented i think - will try and dig out some literature on these.
Can anyone add to/correct my understanding of these?
To cite from the introduction of the upcoming paper on the COST-HOME benchmark dataset:
"Based on the homogenisation of large temperature networks in the Greater Alpine Region and contiguous USA, it is found that the periods between detected inhomogeneities is around 20 years and that the typical size of a break is of the same order as the climatic changes in the 20th century (Auer et al, 2007; Menne et al., 2009). The number of detected inhomogeneities also depends on the algorithms used and the density of the network as smaller inhomogeneities are often not detected. These values are expected to be typical for historical western climate records."
Homogenisation in case the sign of the inhomogeneity is different in summer as in winter does not have to be more difficult. The algorithm of Peter Domonkos, ACMANT, uses annual means and the size of the annual cycle to detect inhomogeneities. I would expect that for such an algorithm the problem is not more difficult. Homogenisation on monthly or seasonal means would also detect such inhomogeneities, although with a bit less sensitivity due to the smaller correlations on shorter temporal averaging scales.
A good review on physical causes of inhomogeneities is still missing, I would say. Some information can be found in Aguilar et al. (2003) and in papers on collocated measurements, e.g. Brunet et al. (2010) or Van der Meulen and Brandsma (2008).
References
Aguilar E., I. Auer, M. Brunet, T.C. Peterson, and J. Wieringa. Guidelines on climate metadata and homogenization. World Meteorological Organization, WMO-TD No. 1186, WCDMP No. 53, Geneva, Switzerland, 55 pp, 2003.
Auer I., Böhm R., Jurkovic A., Lipa W., Orlik A., Potzmann R., et al. HISTALP – Historical Instrumental Climatological Surface Time Series of the Greater Alpine Region. Int. J. Climatol., 27, pp. 17-46. DOI: 10.1002/joc.1377, 2007.
Brunet, M., J. Asin, J. Sigró, M. Banón, F. García, E. Aguilar, J. Esteban Palenzuela, T.C. Peterson and Ph. Jones. The minimization of the screen bias from ancient Western Mediterranean air temperature records: an exploratory statistical analysis. Int. J. Climatol., DOI: 10.1002/joc.2192, 2010.
Menne, M.J., C.N. Williams jr., and R.S. Vose. The U.S. historical climatology network monthly temperature data, version 2. Bull. Am. Meteorol. Soc., 90, no.7, pp. 993-1007, doi: 10.1175/2008BAMS2613.1, 2009.
Meulen, van der, J.P. and T. Brandsma. Thermometer screen intercomparison in De Bilt (The Netherlands), part I: Understanding the weather-dependent temperature differences. Int. J. Climatol., 28, pp. 371-387, 2008.
Post a Comment