Introduction
The Pacific Decadal Oscillation (PDO) is a term coined by Nate Mantua and co-authors in a paper entitled: A Pacific Interdecadal Climate Oscillation with Impacts on Salmon Production in 1997. The PDO is derived from monthly sea surface temperature (SST) anomalies in the North Pacific Ocean, poleward of 20 degrees latitude. Relationships between the PDO and the El Niņo-Southern Oscillation (ENSO) that centers its manifestation in the equatorial Pacific are not presently understood.
In this paper, I will refer to the PDO as the index that is derived from the monthly SST anomalies in the region of the Pacific poleward of 20 degrees latitude.
When one studies the derived PDO over the period of a century, it=s quite convincing that the PDO is a very real cycle (fig 1). The PDO cycle also has a much longer time interval than the ENSO cycle.
While the existence of the PDO seems to be pretty clear, the impacts of the cycle are not as clear, and haven=t been especially well documented for individual climatic regions of the country. The purpose of this paper was to determine impacts the PDO has on precipitation variability in New Mexico.
Dealing with drought is a way of life in the arid Southwest. At least some portion of New Mexico has experienced severe to extreme drought (as described by the Palmer Drought Severity Index) during 56 of the past 102 years. Since the drought of the 1950s, the population of New Mexico has nearly tripled. With the population of New Mexico continuing to grow, the ramifications of drought will become more and more of an issue in the future. This is certainly going to be true of the entire Southwest United States as well as the entire planet. The impact of a A1950s@ type drought will be much greater the next time it occurs. As bad as the 1950s drought was, it still pales in comparison to the drought of the late 1500s. The next time the arid Southwest experiences a drought similar to the late 1500s, if the population does not have some great contingency plan in place, it will be time for millions of people to get up and move. During the next major drought, we may have an impressive array of multi-megahertz or multi-gigahertz computers that are not going to mean a thing. Unless we have an arsenal of desalination plants, and machines capable of drawing air in, separating the oxygen and hydrogen and producing a stream of water, we won=t have the technology to survive in the desert.
Figure 1
Methodology
The relationship between the PDO and New Mexico precipitation was studied in several ways. Data sets for the PDO (as well as figure 1) were obtained from the University of Washington web page. An annual PDO average was calculated for each year from 1931 through 1997. Average precipitation for each of eight climate divisions (fig 2, below) was calculated, for each of those years.
Figure 2
Years were determined for which the PDO varied substantially from zero. In this paper, the PDO was considered to be negative (-) for years in which the average of the monthly index values was less than -.50. Years for which the index was below -1.00 were considered to be Astrongly (- -) negative.@ The PDO was considered to be positive (+) for years in which the average of the monthly index values was greater than +.50. Years for which the index was greater than +1.00 were considered to be Astrongly (++) positive.@
For the sixty-nine years included in the study, 19 years were determined to be either Apositive@ or Astrongly positive.@ Of those 19 years, 9 were determined to be Astrongly positive.@ Eighteen of the years were either Anegative@ or Astrongly negative,@ with 7 of them Astrongly negative.@
An average annual precipitation was computed for each climate division based on data obtained from the National Weather Service=s cooperative observer network, and any aviation observation stations with long-term records. Approximately 175 stations were used in these calculations. Precipitation was considered to be normal for years in which the average for the division was within 10 percent of the long-term average. The remainder of the years were classified according to "above-normal" or "below-normal" status. This was also done for a "statewide" composite of all the climate divisions, as well as for the City of Albuquerque, where approximately one-third of the state=s population lives.
I attempted to determine the likelihood of wet and dry years relative to the PDO cycle, as well as the quantity of precipitation that falls during the negative (cold) and positive (warm) portions of the cycle.
Additionally, precipitation for an entire cycle was considered by looking at the 1944-1976 period as the "cold" side of the cycle in which the PDO was predominantly negative, and the 1977-1997 period that should represent the warm side of the cycle.
Calculated Data
| Div 1 | Div 2 | Div 3 | Div 4 | Div 5 | Div 6 | Div 7 | Div 8 | State | Albuquerque | |
| Wet | 15 | 09 | 24 | 12 | 12 | 18 | 21 | 15 | 15 | 27 |
| Dry | 55 | 48 | 48 | 55 | 52 | 52 | 58 | 58 | 55 | 45 |
| Normal | 30 | 42 | 27 | 33 | 36 | 30 | 21 | 27 | 30 | 27 |
| Div 1 | Div 2 | Div 3 | Div 4 | Div 5 | Div 6 | Div 7 | Div 8 | State | Albuquerque | |
| Wet | 62 | 33 | 48 | 67 | 52 | 48 | 43 | 48 | 43 | 57 |
| Dry | 14 | 05 | 14 | 19 | 19 | 14 | 24 | 14 | 14 | 14 |
| Normal | 24 | 62 | 38 | 14 | 29 | 38 | 33 | 38 | 43 | 29 |
| Div 1 | Div 2 | Div 3 | Div 4 | Div 5 | Div 6 | Div 7 | Div 8 | State | Albuquerque | |
| Wet | 7 | 4 | 4 | 7 | 4 | 4 | 5 | 4 | 3 | 4 |
| Dry | 1 | 1 | 4 | 0 | 2 | 1 | 3 | 1 | 1 | 2 |
| Normal | 1 | 4 | 1 | 2 | 3 | 4 | 1 | 4 | 5 | 3 |
| Div 1 | Div 2 | Div 3 | Div 4 | Div 5 | Div 6 | Div 7 | Div 8 | State | Albuquerque | |
| Wet | 12 | 7 | 9 | 14 | 9 | 8 | 10 | 9 | 8 | 11 |
| Dry | 3 | 2 | 6 | 2 | 4 | 3 | 4 | 3 | 3 | 3 |
| Normal | 4 | 10 | 4 | 3 | 6 | 8 | 5 | 7 | 8 | 5 |
| Div 1 | Div 2 | Div 3 | Div 4 | Div 5 | Div 6 | Div 7 | Div 8 | State | Albuquerque | |
| Wet | 0 | 0 | 1 | 0 | 0 | 0 | 1 | 1 | 0 | 0 |
| Dry | 4 | 4 | 4 | 3 | 3 | 4 | 3 | 4 | 3 | 4 |
| Normal | 3 | 3 | 2 | 4 | 4 | 3 | 3 | 2 | 4 | 3 |
| Div 1 | Div 2 | Div 3 | Div 4 | Div 5 | Div 6 | Div 7 | Div 8 | State | Albuquerque | |
| Wet | 1 | 0 | 3 | 2 | 2 | 1 | 2 | 2 | 1 | 3 |
| Dry | 8 | 9 | 8 | 9 | 9 | 10 | 12 | 12 | 10 | 8 |
| Normal | 9 | 9 | 7 | 7 | 7 | 7 | 4 | 4 | 7 | 7 |
| Category | Total years wet | Total years dry | Total years normal |
| ++ | 39 (54) | 13 (18) | 20 (28) |
| + and ++ | 78 (51) | 27 (18) | 47 (31) |
| (- -) | 03 (05) | 29 (52) | 24 (43) |
| ( - ) and (- -) | 13 (09) | 77 (51) | 54 (38) |
| Category | Div 1 | Div 2 | Div 3 | Div 4 | Div 5 | Div 6 | Div 7 | Div 8 | State | Albuquerque |
| Normal | 11.15 | 16.32 | 15.61 | 13.57 | 9.54 | 17.16 | 13.90 | 11.17 | 13.55 | 8.50 |
| Ave ++ | 13.82 | 18.51 | 17.83 | 17.14 | 11.23 | 19.70 | 17.41 | 13.15 | 16.06 | 10.21 |
| Ave - - | 8.62 | 13.82 | 14.16 | 11.07 | 7.57 | 14.78 | 12.38 | 9.48 | 11.48 | 6.36 |
| Ave(+/++) | 12.27 | 16.62 | 16.60 | 15.33 | 10.39 | 19.11 | 15.76 | 12.66 | 15.48 | 10.16 |
| Ave (-/- -) | 9.74 | 14.50 | 14.31 | 12.09 | 8.20 | 15.06 | 11.50 | 9.72 | 11.75 | 7.61 |
| Category | Div 1 | Div 2 | Div 3 | Div 4 | Div 5 | Div 6 | Div 7 | Div 8 | State | Albuquerque |
| Normal | 100 | 100 | 100 | 100 | 100 | 100 | 100 | 100 | 100 | 100 |
| Ave ++ | 124 | 113 | 114 | 126 | 118 | 115 | 125 | 118 | 119 | 120 |
| Ave - - | 77 | 85 | 91 | 82 | 79 | 86 | 89 | 85 | 85 | 75 |
| Dry/Wet ratio | 62 | 75 | 79 | 65 | 67 | 75 | 71 | 72 | 71 | 62 |
| Ave(+/++) | 110 | 102 | 106 | 113 | 109 | 111 | 113 | 113 | 114 | 120 |
| Ave (-/- -) | 87 | 89 | 92 | 89 | 86 | 88 | 83 | 87 | 87 | 90 |
| Dry to Wet | 79 | 87 | 86 | 79 | 79 | 79 | 73 | 77 | 76 | 75 |
Discussion/Conclusions
There is a strong relationship between the Pacific Decadal Oscillation and precipitation in New Mexico. From tables 1 and 2, the dry years outnumbered wet years nearly four to one (55 to 15 percent of the years) during the last cold phase of the cycle (1944-1976). During the warm side of the cycle (1977-1997), wet years outnumbered dry ones three to one (43 to 14 percent).
When looking at a collection of years in which the PDO was positive or negative (according to the definition in the introduction), the picture is not a lot different. Table 7 shows that when the precipitation is either above or below normal, dry years outnumbered wet years roughly five to one (51 percent to 9 percent) for negative PDOs, while wet years outnumbered dry years nearly three to one (51 percent to 18 percent) for positive PDOs. In this case, differentiating the positive/negative years from the strongly positive/strongly negative years did not change things appreciably. During the positive phase of the PDO, approximately 30 percent of the years have normal precipitation, while normal precipitation occurs during approximately 40 percent of the negative phase.
When looking at precipitation on the cold side of the cycle for each climate division, all divisions had far more dry years than wet. However, if any division seemed to get off "easier," it was division three. Division three actually had 3 wet years, 7 normal years, and 8 dry years during the negative PDO years. Divisions 7 and 8 (both in southern New Mexico) had the greatest number of dry years, with 12. The state as a whole does not do well at all during the cold years, with 1 wet year, 10 dry ones, and 7 normal years.
Western New Mexico, especially the southwest, seemed to benefit most from the positive PDO years. Division 4 (Southwest New Mexico) had 14 wet years, 3 normal years, and only 2 dry years. This is the same area of the state (along with division 8) that exhibits the strongest ENSO signals, and also the strongest summer monsoon signals. Division 1 (Northwest New Mexico) also benefited with 12 wet years, 4 normal years, and 3 dry years. Division 3 seemed to be least affected by positive PDO years, with 9 wet ones, 4 normal ones, and 6 dry years.
Tables 8 and 9 show how the quantity of precipitation varies with the PDO cycle. The state=s average precipitation was roughly 114 percent of normal during positive PDO years. Greatest anomalies were in the south and west. Division 2 was least affected, with precipitation only 102 percent of normal during the positive PDO years.
Unlike percentage of years above and below normal, quantity of precipitation did vary according to the amplitude of the cycle. Table 9 shows precipitation was greater in every division for (++) versus (+) PDOs. The statewide average for strongly positive PDO years was 120 percent of normal, and the divisions in the south and west averaged roughly one-quarter more precipitation during these years.
Table 8 shows that climate divisions average as much as five or six inches more precipitation during the strongly positive PDO years compared to the strongly negative PDO years. This is a huge volume of water when you consider the small annual precipitation averages in the arid Southwest. This variation is no more apparent anywhere as the City of Albuquerque. During strongly positive PDO years, Albuquerque averages 10.21 inches of precipitation. During the strongly negative years, this average is 6.36 inches, a mere 62 percent of the value during the "good" years.
When looking at the results in table 9, it appears the western portion of New Mexico gets the worst of it during the lean years. The division 1 (Northwest) precipitation average during the strongly negative PDO years was only 62 percent of the strongly positive years. This is the same percentage Albuquerque data exhibits. Division 4, which benefits more than any other division during the strongly positive years, received only 65 percent of that moisture during the strongly negative years. Although my first thought would be this decrease is mainly because of the shutting off of the winter precipitation, it does pose a question as to whether or not the PDO cycle has any identifiable relationship with the Southwest Monsoon.
In conclusion, the PDO cycle has a profound influence on New Mexico=s precipitation. It also appears the negative (cold) side of the cycle is related to longer-term droughts such as the one in the 1950s. So, what else can we do with this information?
Looking at figure 1, I think it would be difficult to draw any conclusions about how much precipitation can be expected in any given year. There are anomalies such as 1934 and 1936, when the PDO was strongly positive but precipitation was well below normal. During the 1990s, New Mexico experienced a short but intense drought in 1995 and the first half of 1996, a time when the PDO was also positive. Looking at the graph (fig 1), some obvious questions might focus on why the amplitude of the cold part of the cycle was much greater the 1950s and 1960s compared to what should have been the cold side of the cycle prior to 1923. Is this a trend in which amplitude will continue to increase? What caused the cycle to go negative for three years between 1989 and 1991 during the warm phase? Why was the warm side of the cycle so wet during the 1980s and 1990s while it was dry during the 1930s?
Some of the other anomalies are interesting for their unique contributions. The largest annual average PDO was 1.994 in 1941. This was absolutely the wettest year of the 20th century everywhere in New Mexico. Flooding was extensive and mud slides were common. If you are ever in Chaco Canyon, look at the massive rock slide remnants near the walking trail on the side of the canyon. That happened in 1941.
The absolute lowest annual PDO average of the century was -1.948 in 1955. This value changed very little in 1956 (-1.804), providing the largest negative "two-year" anomaly of the 20th century. By nearly all standards, 1956 was the driest year of the century in New Mexico. Yet there is 1934, with an average PDO of +1.183, the fourth driest year of the 20th century in New Mexico.
In spite of the limitation of being able to use the PDO to forecast a year=s precipitation, perhaps we just need to adjust our scale of forecasting when we are referring to the PDO. To use the PDO, it appears that one year is just too short of a period. The value in the PDO is knowing when the positive and negative portions of the cycle are going to be, and what the ramifications are. Consequently, the PDO may be an affective long-range planning (forecast) tool.
Looking at figure 1, it appears the warm part of the PDO cycle that began in 1977 has about ended. In fact, PDO data since 1997 suggests the PDO probably entered the negative (cold) part of the cycle in 1998. The data suggests dry years will outnumber the wet ones four or five to one. That does not mean every year will be dry, in fact, historical data strongly suggests that won=t happen at all. The data would suggest, during an average 10-year period during the negative part of the cycle, there would be 5 dry years, one wet year, and four normal years. However, when we get to the year 2025 and look back at the last 25 years of precipitation, it is likely we will see averages that are only 60 to 80 percent of the 1977-1997 averages. If we see the PDO decrease to values less than -1.00 for several years, it is likely we will be in a very significant drought similar to the one of the 1950s. However, even if that doesn't occur, the reduced precipitation during the negative part of the cycle will have long-term affects. Decreased precipitation over these years will mean lowered reservoir levels, especially if the main decrease in precipitation occurs during the winter.
There are certainly some caveats in this entire process of extrapolating history forward. One "biggie" might well be global warming. If the globe is warming, all bets are off. Some global warming models suggest New Mexico will continue to get wetter as the globe heats up. However, considering the ramifications and huge impacts a drought similar to the 1950s (or worse) would have on the Southwest United States, prudent is just too mild a word to use here for the planning and contingency plans that should be implemented.
References:
Liles, C.A., 1991: El Niņo and Albuquerque Precipitation. Southern Region Tech Attachment 91-19.
Liles, C.A., 1996 and 1998: La Niņa and Albuquerque Precipitation. Locally distributed.
Liles, C.A., 1998: Monthly and Seasonal Precipitation Distribution in New Mexico. 38 pp. Locally distributed.
Mantua, N.J. and S.R. Hare, Y. Zhang, J.M. Wallace, and R.C. Francis, 1997: A Pacific Interdecadal Climate Oscillation with Impacts on Salmon Production. Bulletin of the American Meteorological Society, 78, pp. 1069-1079.
University of Washington web page: Phillip Mote, 2000: Pacific Decadal Oscillation.
University of Washington web page: UKMO Historical SST Data set for 1900-1981.
University of Washington web page: Reynold=s Optimally Interpolated SST since January 1982.
