The wide expanse of the tropical Pacific Ocean is known by many yet populated by few. Here the El Niño–Southern Oscillation (ENSO) causes fluctuating ocean temperatures that continuously alter global weather patterns, making this geographical location important in the global teleconnections of weather and climate.
During strong El Niño phases, a large-scale eastward shift of anomalous warm sea surface temperatures is accompanied by a large shift in the locations of tropical thunderstorms and hence rainfall. Following the onset of a strong El Niño, many Pacific island nations experience severe drought episodes. Because many island nations are dependent on seasonal precipitation as a vital source of fresh water [Kruk et al., 2015], this shift can cause prolonged water shortages.
For these reasons, local forecasters and water resource managers have asked for tools to assess the probability of wet or dry conditions for a particular area in relation to a specific phase of ENSO. A new satellite-derived visual climate record addresses this need by mapping past precipitation changes in the tropical Pacific in relation to changing phases of ENSO.
The new climate record is a useful tool for water resource managers in the U.S. Affiliated Pacific Islands (USAPI). What’s more, the project highlights the importance of satellite data and their use in the analysis of long-term precipitation trends, particularly those related to ENSO.
A Spatial Problem
Different phases of ENSO have varied effects on the local climate of Pacific island nations and on the global climate as a whole. Yet effects of strong El Niño phases on local Pacific island climates are less widely known than, say, effects on the continental United States. This leaves island nations, buffeted by advancing and retreating ENSO cycles, with a heavy burden of uncertainty.
Developing tools required for assessing climatological variables in the tropical Pacific does not come without challenges. The vast area of the tropical Pacific, with its large spatial gaps between on-site observations, makes it difficult to visualize climate data effectively. Almost all long-term monitoring stations are located on islands that are often separated by large stretches of ocean.
To effectively evaluate climate variables in this region, other sources of data must also be examined in conjunction with in situ data. Data derived from satellites have proven indispensable in meeting these challenges.
In Search of Better Seasonal Precipitation Outlooks
The National Oceanic and Atmospheric Administration’s (NOAA) Pacific ENSO Applications Climate (PEAC) Center currently provides information products based on the ENSO climate cycle for the USAPI, including products that facilitate water conservation. Forecasters issue seasonal rainfall outlooks on a monthly basis using a blend of current observations, dynamic and statistical atmospheric model output, and local expertise [Schroeder et al., 2012].
Forecasters at PEAC currently use a climate record of ENSO-influenced rainfall based on observations from 1955 to 1998 for 66 stations located throughout the Pacific Basin [He et al., 1998]. This in situ–based record provides forecasters with knowledge of specific quantitative precipitation amounts for a given station location in relation to past phases of ENSO.
However, in addition to a lack of spatial coverage, point sources of quantitative rainfall amounts communicate only part of the story. To help facilitate actions in relation to water conservation, we need visual representation of past precipitation during ENSO phases. This visual representation could be used to educate water resource managers and Pacific Islanders about the potential ENSO-related precipitation effects on not just their local islands but the surrounding region as well.
Challenges Across the Pacific Basin
The six areas of the USAPI are shown in Figure 1, along with the USAPI’s exclusive economic zones (EEZs). These zones stretch 370 kilometers from the coast into the surrounding ocean and encompass an area of more than 9 million square kilometers.
To understand precipitation trends over climatological time scales, station records must extend over at least a 30-year period. Only 11 stations in the six EEZs meet this criterion for the time period 1985–2015.
Figure 1 also shows the six in situ station locations within the Republic of the Marshall Islands EEZ that have at least a 20-year record of precipitation and at least a 75% complete record of data—meaning that at least 15 years experienced no data collection hiatuses—from 1985 to 2014. The Marshal Islands EEZ has one station per 300,000 square kilometers—an extremely low density.
In situ station density on time scales shorter than 30 years is sufficient for areas such as the Hawaiian EEZ, where numerous stations cover the highly populated islands. However, in more remote areas of the Pacific, where in situ station data are less prevalent, many long-term station locations are concentrated only near the country capitals. This leaves large gaps on a regional scale.
A Clearer Image
To meet the growing need for better spatial coverage of climate variability within seasonal dry and wet periods, we constructed a record of anomalous precipitation patterns for Hawaii and the USAPI using the publicly available Precipitation Estimation from Remotely Sensed Information using Artificial Neural Networks (PERSIANN) Climate Data Record (CDR). The PERSIANN-CDR’s strength lies in its ability to depict spatial patterns of precipitation, and it has already been used successfully in other data-sparse regions [Miao et al., 2015].
The record was created using the PERSIANN algorithm and GridSat-B1 infrared satellite data. The satellite data are adjusted using in situ gauge data from the Global Precipitation Climatology Project (GPCP). The PERSIANN-CDR product offers global daily coverage of precipitation from 1983 to the present at 0.25° (about 750-square-kilometer) spatial resolution [Ashouri et al., 2015].
We performed a verification analysis by comparing the annual and monthly precipitation amounts calculated by PERSIANN-CDR with data from stations in the Global Historical Climatology Network (GHCN)—an integrated database of quality-controlled observations from land-based stations across the globe. Results indicate that although the PERSIANN-CDR rainfall accuracy varied with different stations, the PERSIANN-CDR accurately depicted seasonal trends of rainfall.
Using the PERSIANN-CDR, we mapped precipitation trends for an area encompassing the USAPI in relation to five defined phases of ENSO for a 30-year period from 1 January 1985 through 31 December 2014. We depicted the seasonal rainfall trends using maps that show the percentage of departure from normal for different areas within Hawaii and the USAPI. We created anomalous precipitation maps to illustrate the past response of precipitation changes in relation to a specific phase of ENSO (Figure 1).
The full downloadable atlas provides a list of specific ENSO events in each phase and the results of the verification analysis.
A Multiple-Source Approach
Grasping the full picture of precipitation patterns involves analysis of visual as well as quantitative sources. Using both the in situ station climate records and the visual climate data products derived from the PERSIANN-CDR provides insight into past precipitation changes associated with ENSO.
To this end, we hope that this visual climate representation will be used as a reference tool to aid water resource managers in Pacific islands. Armed with this atlas, resource managers can, for example, better prepare for forecasted drought episodes by managing water use before strong El Niño phases shift precipitation patterns. In doing so, they will be building more resilient communities.
We thank Lance Watkins, Carl Noblitt, Alejandro Ludert, Carl Schreck, Olivier Pratt, Mark Lander, Luke He, Kevin Kodama, Charles Guard, Thomas Schroeder, Maria Ngemaes, and the NASA DEVELOP National Program Office for their support, edits, and commentary that has strengthened this research. This material is based upon work supported by NASA and NOAA through contract NNL11AA00B and cooperative agreement NNX14AB60A.