About this Project    Numerous studies have documented significant increases in both the frequency and magnitude of extreme precipitation in the northeastern U.S.
since the mid-to-late 20th century. The most recent assessment from the Intergovernmental Panel on Climate Change (IPCC) suggests that the frequency and magnitude
of extreme precipitation in this region will likely continue to increase throughout the 21st century. Such changes could greatly exacerbate the societal impacts of extreme precipitation
in the future. In consideration of these impacts, the Northeast Regional Climate Center (NRCC) has partnered with the New York State Energy Research and Development Authority (NYSERDA)
to downscale global climate model output and create extreme precipitation projections that will ultimately be incorporated into climate change adaptation planning for New York State. Read more...
About this Project    Numerous studies have documented significant increases in both the frequency and magnitude of extreme precipitation in the northeastern U.S.
since the mid-to-late 20th century. The most recent assessment from the Intergovernmental Panel on Climate Change (IPCC) suggests that the frequency and magnitude
of extreme precipitation in this region will likely continue to increase throughout the 21st century. Such changes could greatly exacerbate the societal impacts of extreme precipitation
in the future. In consideration of these impacts, the Northeast Regional Climate Center (NRCC) has partnered with the New York State Energy Research and Development Authority (NYSERDA)
to downscale global climate model output and create extreme precipitation projections that will ultimately be incorporated into climate change adaptation planning for New York State.
Daily precipitation amounts from 157 weather stations in New York and portions of adjacent states and Canada were used to construct partial duration series (the 29 largest observations
at each station during one historical (1970–1999) and three future (2010–2039, 2040–2069, 2070–2099) periods. Precipitation thresholds corresponding to 2-, 5-, 10-, 25-, 50-, and 100-year
return periods were calculated using a regionalized L-moments approach for extreme value fitting. This approach first groups stations together based on similarities in their
extreme precipitation distributions, and then applies L-moments regional frequency analysis to fit a generalized extreme value (GEV) distribution to each station’s partial duration series.
Confidence intervals for the observed precipitation extremes were estimated by randomly selecting (with replacement) 29 precipitation amounts from each station’s historical partial duration
series 1000 times., The 5th and 95th percentile values from these 1000 simulations were chosen to represent the 90% confidence intervals of historical return period precipitation thresholds.
Sub-daily (1-, 2-, 3-, 6-, 12-, 18-, and 24-h) precipitation thresholds were estimated by applying conventional empirical adjustment factors the daily precipitation thresholds. These adjustments have shown
no change in the long-term observed climate record. Since climate model projections of precipitation data at sub-daily resolution are uncertain, any potential changes in the adjustment factors have not
been taken into account in developing projected IDF curves. Users should be aware of this assumption in applications that depend on sub-daily rainfall.
Three different methods were used to downscale future daily precipitation extremes at each station under two IPCC climate change scenarios (RCP4.5 and RCP8.5). The first method employs
quantile–quantile mapping to bias correct areally adjusted precipitation extremes obtained from dynamically downscaled climate model simulations. These simulations consist of regional climate
models (RCMs) run at 50-km resolution and driven by atmosphere–ocean general circulation models (AOGCMs) from Phase 5 of the Couple Model Intercomparison Project (CMIP5). The second method, a
variation of the delta method, computes differences in simulated precipitation extremes between CMIP5 future and historical periods, and applies these differences toward observed precipitation
extremes. The third method combines quantile–quantile mapping with a unique approach for downscaling daily precipitation extremes from historical analogs. This analog approach involves a multi-step
procedure in which the occurrence of extreme precipitation on a given CMIP5 model day is first predicted based on the observed probability of extreme precipitation on that day’s closest historical
analog days. Then, if extreme precipitation occurred on the selected analog day(s), the precipitation observations associated with the historical analog day(s) are used to ascribe precipitation
amounts on the corresponding model day. Across all three downscaling methods, 49 unique sets of extreme precipitation projections were generated for each climate scenario–time period combination.
These 49 simulations form an ensemble of projections that is summarized by an ensemble mean, and the ensemble member corresponding to the 10th and 90th percentiles of the 49 simulations.
As one of the main project deliverables, this webpage showcases final research products and provides a detailed summary of the project methodology. Publicly available research products
include: 1) historical and future downscaled intensity-duration-frequency (IDF) curves for each station, 2) gridded maps illustrating projected changes in return period precipitation amounts,
and 3) gridded maps illustrating projected changes in the recurrence intervals of historical precipitation thresholds. Smoothed IDF curves were constructed by calculating precipitation intensities
for the sub-daily precipitation thresholds and fitting a logarithmic regression to the intensity-duration relationships. The gridded percent change maps were created by computing percent
changes in precipitation thresholds between historical and future periods at each station, and interpolating these station values to 0.5° × 0.5° grid cell values. The gridded recurrence
interval maps were created by estimating the future recurrence intervals of historical precipitation thresholds at each station and interpolating these station values to 0.5° × 0.5° grid cell values.