Introduction

Precipitation occurs directly on to the surface of Long Island Sound and its watershed as rain and snow. It can impact ecosystems in a variety of ways. The over-land precipitation leads to runoff to streams, ground water, and marshes where it can impact the turbidity, the nutrient level and the salinity. The mechanisms though which precipitation acts as an ecological driver can be complex. For example, high flow rates in streams that arise during high precipitation rate events can increase sediment concentration though bank erosion and stream migration. In addition, that atmospheric deposition of particulate nitrogen to the earth and surface of Long Island Sound is rapid when rain begins. On an annual basis, the total fluxes are likely influenced by the number of rain events or days with rain.

An increase in the frequency of extreme rainfall events in northeastern United States has been suggested by the analysis of DeGaetano (2009) and USGCRP (2009). These studies focused on the probability of the occurrence of very large (e.g., 2 inch/day) events and showed that in New England these were more frequent now than in 1960. The predictions of climate models for the effect of recent warming on large storms of 2013 was less clear (Knutson et al., 2014). The intrinsic variability of the climate systems could account for much of the current observations. The analysis on future changes predicted by models reported in Walsh et al. (2014) and Horton et al. (2014) concluded that though the magnitude of the change of mean annual precipitation amounts was very uncertain, there was high confidence that the magnitude of annual daily maximums would increase.

Precipitation Observations Measurements of precipitation are archived at NODC and can be accessed through the “Climate Data Online” web site. Figure 1 shows the locations relative to the shoreline of the Sound	Figure 1. Site map of meteorological stations. Upland CT stations are green, coastal CT stations are blue, and LI stations are red.
Trend and Variability Analysis Figure 2 shows the mean and  a standard error of the total precipitation at each of the stations with reliable data. Note that the standard error was computed without correction for serial correlation, and therefore might be slightly underestimated. The values are appreciably higher than the estimate of Miller et al. (2002) of 44.8 inches which included many more stations to the north and none from Long Island, and omitted contributions from snow. These data are most relevant to the coastal habitats of Long Island Sound.	Figure 2. Comparison of long term mean annual precipitation. The red symbols show the statistics for the stations on Long Island while the blue and green show the data for coastal and upland Connecticut, respectively.
Figure 3. Average by year of all annual precipitation estimates across the three geographic areas.	Figure 4. Regional mean precipitation filtered with a box-car moving average (black line) and  long term trend estimated using un-weighted linear regression and all the data (blue dashed line). The slope shows an increase of 0.5 inches per century. The green dashed line also shows a regression based on data between 1915 and 1994. The rate of increase is more rapid at 5.1 inches per century.
The differences between the trends in the three sub-regions are computed by subtracting the mean of all the stations (Figure 4) from the mean of each of the other three series. Figure 5 shows the results after a 7-year running-mean filter has been applied to each series. The green line shows that the annual precipitation anomaly in the inland areas since 1935 is generally higher than the others in keeping with the results in Figure 2. The coastal Connecticut stations (blue) and Long Island Stations (red) show a generally decreasing trend. The decadal-scale variability in the upland CT (green line) is negatively correlated with the variations on Long Island and of approximately equal magnitude (about 5 inches). A significant peak in the cross correlation of these series occurs at zero lag with a magnitude of -0.5. This is consistent with the conclusion that the regional mean is not changing very much but the Long Island and coastal area have experienced a reduction in the annual precipitation while the inland areas have experienced a slight increase.	Figure 5. Difference between the box-car filtered average precipitation across the entire region and the filtered average in the three sub-areas.
Figure 6 displays the variation in the number of days on which greater than zero rainfall and snow were reported. The analysis reveals that in most years it rains approximately 10 fewer days on Long Island (mean of 100 days) than in coastal Connecticut (120 days). The area wide mean over the whole record, and in inland Connecticut, is 115 days with a standard deviation of 13 days. Coastal Connecticut endured more rain days in the period 1925-35 but there was a decline until 1960 and since then the variation has been less. Long Island also endures fewer days of snow (11) than Connecticut (14) in most years. However, the number of snow days has been declining throughout the 20th century across the region. This is particularly apparent on Long Island where the mean of the last 30 years shows 3 days less snow than the earlier period.	Figure 6. a) Evolution of the average across the three region of the number of days in which rain was measured. A seven year box-car filter has been applied to the records. b) Corresponding data for snowfall. The red lines show the Long Island station data and the blue and green show the coastal and inland Connecticut stations.
Figure 7 shows how the fraction of observations greater than 2 inches per year varies with time. The percentage of observations in each sub area is plotted using the same color code and added to the other sub areas. Note that 5 year intervals were used to stabilize the estimates but the percentage of observations per year is presented. The distribution in the three areas is highly correlated and of comparable magnitude. There appears to have been a large increase in the occurrence of high rainfall days after 1940 and then a dip in the mid-1960s. Between 1975 and 2000, the frequency of high precipitation reached a maximum and then declined. This is not completely inconsistent with the analysis of DeGaetano (2009). His conclusion applies to a much larger region and based on extreme value statistics using only data before 2007. Our results show no evidence of an increase in extreme precipitation events in the coastal areas surrounding Long Island Sound.	Figure 7. Percentage of all observations in excess of 2 inches per day in 0.25 in bins. Red shows Long Island, and blue and green show coastal and inland Connecticut.

Summary

We have identified a subset of the NODC weather observation stations that span the area surrounding Long Island Sound and evaluated the data quality for the estimation of Total Precipitation (rain and snow). The station data was analyzed to discriminate differences from coastal Connecticut, inland Connecticut and Long Island. We find that the mean annual precipitation at the inland stations is significantly higher than in the coastal areas and Long Island. The records show substantial year to year variability but that the area-wide average precipitation does not show a significant long term trend. There is, however, long term trends within the region with the annual precipitation in the coastal regions negatively correlated at decadal scales with those inland. The inland stations also show an increasing trend. The number of days of rain and snow were also examined and we show that there has been a slight decline in the number of snow days on Long Island over the last century. The number of days with rainfall exhibits decadal scale cycles but has not appreciable changed on Long Island or inland Connecticut, although there is evidence of fewer rain days in coastal Connecticut in the recent past relative to the 1930s. Finally, we examined the frequency of occurrence of rainfall events leading to more than 2 inches of rain in a day. We find no evidence of an increasing trend the in the frequency of these events.

References

DeGaetano, A.T. (2009). Time-Dependent Changes in Extreme Precipitation Return-Period Amounts in the Continental United States. Journal of Applied Meteorology & Climatology, vol. 48, no. 10, pp. 2086–2099.

Horton, R., G. Yohe, W. Easterling, R. Kates, M. Ruth, E. Sussman, A. Whelchel, D. Wolfe, and F. Lipschultz (2014) Ch. 16: Northeast Climate Change Impacts in the United States: The Third National Climate Assessment, J. M. Melillo, Terese (T.C.) Richmond, and G. W. Yohe, Eds., U.S. Global Change Research Program, 16-1-nn.

Knutson, Thomas R., F. Zeng, and A. T. Wittenberg  (2014) Seasonal and Annual Mean Precipitation Extremes Occurring During 2013: A U.S. Focused Analysis [in "Explaining Extremes of 2013 from a Climate Perspective"]. Bulletin of the American Meteorological Society, 95(9), S19-S23.

Miller, D. G. S. Warner, F. L. Ogden and A.T. DeGaetano (2002) Precipitation in Connecticut. CT. Inst. Water Resources, Special Report. (Available at http://www.ctiwr.uconn.edu/PrecipinCT/precip.pdf).

U.S. Global Change Research Program (USGCRP) (2009) Global Climate Change Impacts in the United States, Thomas R. Karl, Jerry M. Melillo, and Thomas C. Peterson, Eds. Cambridge University Press.

Walsh, J., D. Wuebbles, K. Hayhoe, J. Kossin, K. Kunkel, G. Stephens, P. Thorne, R. Vose, M. Wehner, J. Willis, D. Anderson, S. Doney, R. Feely, P. Hennon, V. Kharin, T. Knutson, F. Landerer, T. Lenton, J. Kennedy, and R. Somerville  (2014) Ch. 2: Our Changing Climate. Climate Change Impacts in the United States: The Third National Climate Assessment, J. M. Melillo, Terese (T.C.) Richmond, and G. W. Yohe, Eds., U.S. Global Change Research Program, 19-67. doi:10.7930/J0KW5CXT.

Back to top