Confidence (Continued)

Over the past week, I have been continuing the confidence analysis which was introduced in last week’s post. This analysis is proving to be extremely worth-while, allowing me to both assign a level of confidence to each Extreme Climate Index as well as look at the spatial variation of each trend.

The Data Sets

Because extreme climate trends can be complex and vary over small areas, we wanted to look at more than just one meteorological time series. Figure 1 is a “snapshot” overview of the different data products being used in the analysis. (And I must emphasize that everything we do is anchored at the CBNERRS reserves; you  can read about our correlations here and here.)

Figure 1: A conceptual time line of the different meteorological datasets being used in this study.

Figure 1: A conceptual timeline of the different meteorological data sets being used in this study.

For our climate analysis, we are using two gridded products: HadEX2 and GHCNdex. The HadEX2 is a 3.75° by 2.5° grid, meaning that it is an area-weighted average for each Extreme Climate Index. The HadEX2 incorporates hundreds of individual weather stations, all at least 11 years in length, allowing us to capture a regional picture of Chesapeake Bay’s extreme climate patterns and trends. The length of the HadEX2 is from 1901-2010, so we can retrospectively reconstruct 110 years of historical weather trends.

We also are using the GHCNdex gridded product for a generalized comparison. The GHCNdex uses smaller grid cells, 2.5° by 2.5°, providing a more tightly confined regional average. However, this gridded product is shorter, starting at 1951, and contains fewer weather stations.

Additionally, we have included 18 individual weather stations from the National Climatic Data Center (Global Historical Climate Network-Daily). Each station starts at a different time (ranging from 1892 to 1955) and comes we a need for extensive quality control. As a comparison, the HadEX2 product already includes some quality control efforts.

Thus, for each Extreme Climate Index, we have 20 different data sets. (That means that we will be assessing 520 climate trends in this effort!). The end result: when we observe a trend in the HadEX2 gridded data product, we can assign how confident we are in that trend represents the entire region’s spatial heterogeneity (or the weather variation between different locations within our HadEX2 grid).

Assigning Levels of Confidence

For this section, I am going to give a glimpse into how I have been organizing this data. I absolutely love color mapping, as I think it visually shows trends and patterns exceptionally well. As Table 1 shows, I have been shading all positive trends in green and negative trends in orange while leaving the data sets with no statistically significant trends blank.

Table 1: A color-shading example for the 21-year mean and variability trends for the Growing Season Length. Note that no shading indicates that there was no significant trend.

Table 1: A color-shading example for the 21-year mean and variability trends for the Growing Season Length. Note that no shading indicates that there was no significant trend.

Hopefully, it is easy to see the fairly consistent mean increase trend for the growing season length. This analysis predicts that 85% of our data sets had a significantly positive 21-year mean trend. Alternatively, the 21-year variability analysis is much more complex! While 65% of the data products had no observed trend in variability, 3 data sets had a positive increase in variability while 4 had a decrease in variability.

Thus, we can be fairly confidence that the growing season length is increasing throughout the near-shore Chesapeake Bay region. By IPCC terminology, this would be termed as a likely trend. However, we cannot be confident that any trend in variability exists for this specific region.

Climate Normals

Table 2:

Table 2: The climate normal means for the growing season length from 1971-2000 and 1981-2010.

As part of my efforts for the confidence analysis, I have also been calculating the 30 year climate normals for all 520 weather products; it’s overwhelming, but necessary! These climate normal are 30 year averages, which remove inter-annual and some decadal variability, proving a solid baseline for what the current mean value is for each Extreme Climate Index.

I had been calculating this baseline following the IPCC 5th Assessment Report from 1971-2000. During the Think Tank meeting, there was also a request to calculate the new climate normal from 1981-2010 as well.

These climate normal means represent the present baseline for the Extreme Climate Indices, such as Growing Season Length (Table 2).

More work must be done, however, this climate normal analysis and confidence investigation is almost complete!

Kari Pohl

About Kari Pohl

I am a post-doctoral researcher at NOAA and the University of Maryland (Center for Environmental Science at Horn Point Laboratory). My work investigates how climate variability and extremes affect the diverse ecosystems in Chesapeake Bay. I received a Ph.D. in oceanography from the University of Rhode Island (2014) and received a B.S. in Environmental Science and a B.A. in Chemistry from Roger Williams University (2009). When I am not busy being a scientist, my hobbies include running, watching (and often yelling at) the Boston Bruins, and taking photos of my cat.
This entry was posted in Data and tagged , , . Bookmark the permalink.

Leave a Reply

Your email address will not be published. Required fields are marked *