Coastal Ocean Analytics

Detecting Climate Change Impacts In Long Island Sound

River Discharge
Sea Level
Air Temperature
Water Temperature
Download Data

Problem Definition/Background

The adaptive management of the resources of Long Island Sound (LIS) requires on-going observations to characterize the variability and change of the environment and ecosystem it supports. It is critical that changes that result from local human activities (and can potentially be regulated) be separated from those that are a consequence of natural cycles and global scale processes. Therefore, it is important to examine existing measurements from Long Island Sound and its watershed to determine whether changes that have been observed at the global scale have discernible and important impacts in the region.

The detection of climate change signatures in observations is a very challenging task. The aggregation of temperature measurements from around the world, together with an extensive and sophisticated program of data quality checking, bias corrections, and weighting to correct for heterogeneous sampling density, was central to the development of the first broadly accepted evidence of global warming. The products of these programs have become known as the NASA-GISS (Hansen et al., 1999), the NOAA-NCDC (Reynolds and Smith, 1994) and the Hadley-CRU (Jones et al., 1999) temperature climatologies. Since the implications of global warming are vast, and the costs of mitigating the effects of change are huge, the results of these analyses have been challenged repeatedly in the literature and in the public press.
Among the more reasonable objections were that the data and analysis methods were not independent, the groups shared ideas during the development process, and that the data analyses procedures were not transparent. These criticisms have motivated an entirely new analysis by the Berkley Earth group. They repeated the process of data aggregation, screening, etc., and developed an open source approach to sharing data and analysis software. They recently released preliminary results and have submitted their reports for publication (Rohde et al., 2011). The figure shows a comparison of the decadal average of the global average of land observations for the three earlier analyses and that of the Berkley Earth group. The grey shaded areas are the 95% confidence interval of the Berkley Earth analysis (black line). All four trends are in close agreement since 1950 when instruments and data standards improved. Between 1900 and 1950, the Berkley Earth results are slightly lower than the others. Overall, the case for warming by 1.3șC since 1900 is strong. The project also analyzed regional variations in the change of mean atmospheric temperature since 1960 and reported Bridgeport, CT, had warmed 2.6 ± 0.45șC.
Decadal land surface average temp from Rohde et al., 2011(from Rohde et al., 2011)

 The factors that made detecting change in global average temperature difficult include:

  1.  The magnitude of the change is small compared to the variance due to sample location (latitude, longitude and altitude).
  2.  The magnitude of the change is small compared to the variance at a site due to daily and annual cycles.
  3.  There are long period (decadal) oscillations in the records at many sites.

These challenges would be easily overcome if the distribution of stations was uniform, the sample frequency resolved daily variations, and the observation interval spanned many decades. However, the data set was not perfect. Additional problems include:

  1. Sample locations are dense in some areas, sparse in others, and altogether absent in Antarctica until 1950.
  2. Instrument design and performance, sampling rates and times changed during the measurement interval.
  3. Few sample stations spanned the whole period.

The detection of signatures of climate change in observations from Long Island Sound has to overcome all of these difficulties. In addition, the lack of sampling at a rate that resolves the daily variations in water properties (temperature, salinity, etc.) in the Sound prior to 2004 makes the trend detection even more difficult. Though there is clearly a significant change in the global average air temperature and this clearly influences many other processes in the environment, characteristics of the variability in measurements and inadequacies in the available data records may frustrate our ability to detect unambiguously the changes. It is, therefore, likely that analyses of some types of data will conclude that there is no detectable local climate change signal. Of course, this does not mean that no link exists.

Since there are well established large scale patterns of variation that have decadal-scale periods (e.g., El Nino-Southern Oscillation and the North Atlantic Oscillation), a data record from a limited geographic region that is shorter than four decades is unlikely to yield a credible estimate of a trend unless the decadal-scale cycles can be extracted. Therefore, analysis should be restricted to records that can be aggregated to intervals longer than 40 years.


The Sentinel Monitoring Strategy identified thirty-five sentinel characteristics of the ecosystem that were important and identified an associated index that was measurable. However, few of these had extensive data records and none has been shown to exhibit trends at the global-scale. Most Sentinels were also linked to a subset of ten significant ecological drivers:

  1. precipitation
  2. stream flow
  3. sea level
  4. air temperature
  5. water temperature
  6. salinity
  7. wind (speed and direction)
  8. relative humidity
  9. pH
  10. groundwater levels
There are considerable data in the Long Island Sound region for each of these quantities, much of which is described in the Sentinel Monitoring Program data citation clearinghouse. This project focuses on collating and analyzing these ecological driver variables since the data records are longest and most likely to yield clear results. We also address trends in three Sentinels (Lobster Habitat, Marsh Flooding, and Sea Cliff Erosion) by exploring nonlinear statistics such as the duration in excess of a threshold (e.g. warmer that 20șC or non-tidal water level anomalies greater than 1.4 m) and parametric estimates of significant wave height based on wind observations.

Our specific objectives were:

  1. To identify all available data for each of the variables and aggregate series to synthesize records that are as long as possible.
  2. To analyze data records of longer than 40 years to identify the long term variations and trends. We will then acquire archived indices of global scale atmosphere-ocean cycles and employ correlation analyses to establish what fraction of the long term variations in the ten ecological drivers variables can be explained by the cycles and what can be attributed to climate change.
  3. To analyze the river discharge, wind and temperature records to establish the inter-annual variations in thresholds such as the center of volume flow, frequency of winds from the northeast and southwest in the summer, duration of temperature in excess of thresholds and nonlinear statistics to be chosen after consultation with the LIS Science and Technology Advisory Committee (STAC).
  4. To create a proxy record of significant wave heights based on recent buoy observations and archived coastal wind records.
  5. To disseminate the raw data, the time series resulting from our analyses, and the programs used to create them through a project website.
  6. To provide advice to the LIS Program on future monitoring and analyses that are necessary to better link our products to the Sentinels.


Hansen, J., R. Ruedy, J. Glascoe, and M. Sato (1999). GISS analysis of surface temperature change, J. Geophys. Res., 104(D24), 30,997–31,022, doi:10.1029/1999JD900835.

Jones, P.D., M. New, D.E. Parker, S. Martin and Rigor, I.G. (1999). Surface air temperature and its variations over the last 150 years. Reviews of Geophysics 37, 173-199.

Reynolds, Richard W. and Thomas M. Smith (1994). Improved Global Sea Surface Temperature Analyses Using Optimum Interpolation. J. Climate, 7, 929–948. doi: 10.1175/1520-0442(1994)007.

Rohde, R., R.A. Muller, R. Jacobsen, E. Muller, S. Perlmutter, A. Rosenfeld, J. Wurtele, D. Groom and C. Wickham (2011). A New Estimate of the Average Earth Surface Land Temperature Spanning 1753 to 2011. J. Geophys. Res. (submitted).

Back to top