United States temperature and precipitation observations for the period 1950-2002 were used to obtain High Frequency (HF) and Trend Adjusted (TA) composites by ENSO phase. ENSO events were chosen using a historical sea-surface temperature data set and NOAA definitions of El Niño and La Niña. Years that do not satisfy the NOAA definitions are ENSO-neutral. The gridded precipitation and surface air temperature data sets were developed at NOAA's Climate Prediction Center as described. Results are presented for each 3-month season (i.e. JFM, FMA, MAM, etc). ENSO temperature and precipitation percentages are expressed as departures from random chance (33.3%) of the indicated category (based on a tercile distribution), for El Niño, La Niña and ENSO-neutral events. Comparisons are made to results from Climate Division data for the same period (1950-2002) and for a longer period (1895-2002).
Observed U.S. surface air temperature data are from the analysis of Janowiak et al. 1999. Daily data were gridded to a horizontal resolution of 0.5° lat x 0.5° lon for the period 1950-2002 (June). Observed U.S. precipitation data are from the Unified Precipitation Reanalysis of Higgins et al. 2000 and Shi et al. 2002. Daily data were gridded to a horizontal resolution of 0.25° lat x 0.25° lon for the period 1950-2002 (June). Both temperature anomalies and precipitation anomalies for the basic composites (section IV of the webpage) are defined as departures from base period (1971-2000) mean values. Time series of the seasonal anomalies for each 3-month season (JFM, FMA, MAM, AMJ,...) were generated for the temperature and precipitation fields prior to the composite analysis.
3. High Frequency (HF) and Trend Adjusted (TA) composites
There have been significant trends in precipitation and surface air temperature at many locations in recent decades, so it is worthwhile to examine the influence of trends on ENSO composites. For this purpose, two basic types of composites are examined:
· High-frequency (denoted HF) composites
· Trend adjusted (denoted TA) composites
HF (or detrended) composite anomalies (sections II and III of the web page) are obtained after first removing an 11-year (15-year) running mean from the raw temperature (precipitation) seasonal time series. At the beginning and end of the time series the closest possible approximation to the low-frequency signal is used. For 2000, for example, 1992-2002 is used. The HF composites are then computed directly. The TA composites are obtained after adding the most recent 11-year (15-year) mean value (computed from the raw data after removing base period 1971-2000 means). In this way the HF time series are “adjusted” to account for the most recent estimate of the low-frequency climate signal.
In order to determine the percentage of events that occur in a particular tercile class (expressed as percentage anomalies in sections II and III on the webpage), it is necessary to determine class limits. Tercile class limits for the HF composite are determined from ranked values of the HF time series for the entire period of record. Class limits for the TA composite are determined from the original (raw) time series using the period 1971-2000 fit to a normal (gamma) distribution for temperature (precipitation). The 1971-2000 base period is used so that results can be applied for forecast operations at CPC, where the current base period is 1971-2000. Thus, these HF and TA composites can be viewed as the high-frequency and low-frequency components of the seasonal forecasts.
3.1 Anomalies (High Frequency and Trend Adjusted)
HF and TA composites of U.S. temperature anomalies in °C (precipitation anomalies in mm day-1) for El Niño, La Niña and ENSO-neutral events during 1950-2002 (June) are available in sections II and III of the webpage.
3.2 Percentages (High Frequency and Trend Adjusted)
HF and TA temperature and precipitation percentages, expressed as departures from random chance (33.3%) of the indicated category, for El Niño, La Niña and ENSO-neutral events during 1950-2002 (June) are available in sections II and III of the webpage. Percentages are based on the fraction of the total number of events that occur in the indicated category. Dashed lines are the 1971-2000 climatology in °C.
3.3 Clickable maps
Clickable maps showing the percentage of the total number of El Niño, La Niña, or ENSO-neutral events in each temperature and precipitation tercile for each season (JFM, FMA, MAM,…) are found in sections II and III of the webpage. The clickable maps are subdivided into 2°x2° grid boxes, and results are based on the average of all points within the box and on its edges.
3.4 Temperature and precipitation anomalies (Basic Composites)
Composites of U.S. temperature and precipitation anomalies in °C (precipitation anomalies in mm day-1) for and ENSO-neutral events during 1950-2002 (June) are available in section IV of the webpage. Anomalies are computed with respect to base period 1971-2000 mean values
3.5 ENSO Composites Based on CPC's Consolidation Forecast for Niño 3.4 SST
Seasonal U.S. temperature and precipitation projections can be obtained at lead times out to 12.5 months using CPC's official consolidation forecast for tropical Pacific SST's. In particular, the standardized anomaly version of the consolidated forecast for Niño 3.4 SST's is used to obtain weights (i.e. a projection fraction) for El Niño, La Niña and ENSO-neutral conditions at leads out to 12.5 months. The weights are applied to the El Niño, La Niña and ENSO-neutral composites and summed to obtain the projections at each lead. At early leads the composites are dominated by the HF component of the forecast (i.e. a particular ENSO phase), while at longer leads the trend component (i.e. OCN) dominates. These composites are used as an objective tool for CPC's Official Seasonal Outlooks.
For most values of n, where n is the number of ENSO events in a particular composite, the number of events required to achieve 95% confidence is n/2. This translates to a percentage anomaly of ~17% (i.e. 50%-33.3%). Thus, for simplicity, shaded areas on the maps at and above 20% can be considered significant at the 95% confidence level.
Higgins, R. W., W. Shi and E. Yarosh, 2000: Improved United States precipitation quality control system and analysis. NCEP/Climate Prediction Center Atlas Number 7, 40 pp. [Available from Climate Prediction Center, World Weather Building, Room 605, Camp Springs, MD, 20746].
Janowiak, J. E., G. D. Bell, and M. Chelliah, 1999: A gridded data base of daily temperature maxima and minima for the conterminous United States: 1948-1993. NCEP/Climate Prediction Center ATLAS No. 6, 50 pp. [Available from Climate Prediction Center, World Weather Building, Room 605, Camp Springs, MD, 20746].
Shi, W, R. W. Higgins and E. Yarosh, 2002: A Unified Raingauge Dataset and Multi-year Daily Precipitation Reanalysis for the United States. J. Geophys. Res