LOCUS Tutorial Research Project One:
Influence of El Niño on the Gulf of Panama Seasonal Productivity Cycle

Table of Contents

• 1. Research Setting
• 2. Primary Research Question
• 3. Investigation Plan
• 4. Data Access and Visualization Methods
• 5. Preliminary Analysis
• 6. Refinement of Analysis
• 7. Statement of Results
• 8. Discussion of Results
• 9. Statement of Conclusions
• 10. Questions for Further Investigation

1. Research Setting

The setting for this research project is the Gulf of Panama, located south of the eastern end of the isthmus of Panama. The Gulf of Panama is directly south of the Panama Canal. The western land boundary of the Gulf is the Azuero Peninsula, and the eastern land boundary is the northern coast of Colombia in South America.

The Gulf of Panama has a well-defined seasonal primary productivity cycle. This cycle is caused by the occurrence of high pressure systems in the Gulf of Mexico and Caribbean Sea that occur during the Northern Hemisphere winter. When these high pressure systems enter the Caribbean Sea north of Panama, they create a marked difference in atmospheric pressure between the northern side of isthmus of Central America and the southern side. The high mountains of the Central American cordillera prevent the pressure from "equalizing" over a broad area. As a result, very strong winds occur in the only three gaps in the cordillera: in southern Mexico, in central Nicaragua, and the Panama Canal. These winds, funneled through the gaps, extend for a large distance over the adjacent ocean waters. The winds are so strong that they mix water lying deeper in the oceanic water column with surface waters. The deeper ocean waters have significantly higher nutrient concentrations than the surface waters, so that when the deeper waters are mixed to the surface, phytoplankton utilize the nutrients for markedly increased photosynthetic productivity.

It is easy to perceive the seasonal productivity cycle for this region in chlorophyll concentration data from instruments such as SeaWiFS and MODIS. (In fact, in the Coastal Zone Color Scanner mission data composite on the left side of the LOCUS banner image at the top of this page, this area of productivity can be seen just above the letters "ory" in "Laboratory".) A complicating factor, however, is the presence of heavy cloud cover over the Gulf of Panama during late spring and nearly all of summer (in the Northern Hemisphere), which may introduce uncertainty into the chlorophyll concentration data for those seasons compared to the less cloudy winters.

The images below, created using the data analysis tool GIOVANNI, illustrate the seasonal cycle in the Gulf of Panama during the years 2000-2001. For convenience, the seasons are: winter, December 2000 through February 2001; spring, March 2001 through May 2001; summer, June 2001 through August 2001; and autumn, September 2001 through November 2001. In particular, note the amount of missing data due to cloud cover in the summer image.

2. Primary Research Question

This research project is concerned with the possible influence of the El Niño climate phenomenon on the seasonal productivity cycle in the Gulf of Panama. Data from SeaWiFS covers two El Niño/La Niña events: the strong El Niño event in the years 1997-1998 (with a strong La Niña following it), and the weaker event occurring in 2002-2003. So our primary research question is formulated:

Can an influence of El Niño/La Niña in the Pacific Ocean be perceived in SeaWiFS chlorophyll concentration data for the Gulf of Panama, and if so, how strong is this influence?

3. Investigation Plan

For this investigation, we plan to utilize the SeaWiFS 9 kilometer resolution global data that are available for analysis using GIOVANNI. Use of these data provide the opportunity to examine patterns of chlorophyll concentration, which is an indicator of phytoplankton productivity, over several years. Our study area will be the Gulf of Panama, which lies south of the eastern part of the isthmus of Panama.

We will examine the patterns of productivity for each year, with particular attention regarding the patterns of productivity in 1997-1998, when a strong El Niño/La Niña event occurred, and 2001-2002, when a much weaker El Niño/La Niña event occurred.

4. Data Access and Visualization Methods

The SeaWiFS 9 km chlorophyll data are processed into monthly files containing the average chlorophyll concentration for each 9 x 9 km "square" area over the world's oceans. GIOVANNI accesses these files and provides the capability of selecting areas of interest for examination. GIOVANNI can be used to create a map of the chlorophyll concentrations for the area averaged over selected time intervals, concentration vs. time plots for chlorophyll concentration (averaged over the entire selected area), month-by-month animations of data for the selected area, or Hovmoller plots for the area.

Hovmoller plots can be particularly useful data visualizations to detect variations over time and space. Hovmoller plots display data in a time vs. longitude or time vs. latitude format. Thus, for the same area, a comparison of the spatial variability over time is easy to comprehend.

[Hovmoller plots are named after a Swedish scientist and meteorologist, Ernest Hovmoller, who first demonstrated their usefulness in this type of analysis in a paper published in 1949. Because these data presentations have become so ubiquitous in data analysis, particularly for earth science, Ernest Hovmoller has achieved a remarkable level of scientific fame: it is easy to find references to Hovmoller plots or Hovmoller analysis, but it is quite difficult to find references to Ernest Hovmoller!]

5. Preliminary Analysis

We initially chose the area bounded by these coordinates: northern latitude 10 degrees North, southern latitude 2 degrees North, western longitude 83 degrees West, eastern longitude 78 degrees West. The first plot simply displays the average chlorophyll concentration for the entire period of time, September 1997 to December 2003.

This figure isn't particularly informative. As we have already seen the pattern of annual variability (the four images shown before), we know that the zone of higher productivity is found in a fairly narrow band and during the winter. The average plot above gives a general idea of the area where productivity is increased, but it provides no information about the variability of productivity over time, or over the region of interest.

The next thing to do is to numerically examine the average chlorophyll concentrations by month over the entire time period.

This figure is more interesting. It appears to confirm that chlorophyll concentrations in the Gulf of Panama are much higher during the winter months, which is what was expected. However, the average high concentrations for 1997-1998 and for 2002-2003 are not much different than the average high concentrations for other years.

So now it's time to generate a Hovmoller plot of these data. In this case, a latitude vs. time plot is chosen (it's also possible to generate longitude vs. time plots). The reason for choosing latitude vs. time rather than longitude vs. time is due to the fact that the high productivity occurs in such a narrow north-to-south area. This means that we would expect to see more spatial variability over latitude as compared to longitude. Remember that in this figure, north is to the right, south is to the left.

A quick examination of this Hovmoller plot looks promising. In 1997-1998, the highest chlorophyll concentrations didn't extend nearly as far south as in other years (except for the productive area located at about 2.8 degrees North). A comparison of 2001-2002 to the productive winter seasons in other "normal" years suggests that the concentrations were slightly lower during that period. So there may be a detectable signal of El Niño/La Niña events in these data.

However, there are a few problems that we need to examine first.

6. Refinement of Analysis

The Hovmoller plot indicates that in 1999 and a few other times, there are some missing data (the white patches). The white patches north of 9 degrees North are actually in the Caribbean Sea, so we don't have to consider them. Our first look at the data indicate that summer is a particularly cloudy period; it was so cloudy this particular spring and summer that no data were acquired for small areas in a one or two month period. Is this a problem?

Probably not. When we look at the data for spring and summer for all of the other years except 1998 (more on that later), the concentration range appears similar. So even though some data are missing, there isn't any reason to expect that it was significantly different than when the data were available.

The other problem is that our study area of interest was too large. A southern boundary of 2 degrees North includes a small portion of the coast of Colombia, where chlorophyll concentrations appear elevated (which is characteristic of chlorophyll concentration data in general anywhere near the coast). This coastal region is the band of productivity located at about 2.8 degrees North. So we need to move our southern boundary northward to about 3.5 degrees North. Here is the new Hovmoller plot:

This plot is much better. It is still distracting, however, to include the coastal area between 8 and 9 degrees North, where there is little apparent variability in productivity. So now let's restrict the Hovmoller plot to between 7 degrees North and 3.5 degrees North.

Now we have a very good plot highlighting the most variable region of the Gulf of Panama. It is clear in this plot that the 1997-1998 winter, during the strong El Niño, was much less productive than the "normal" winter pattern. And it is also fairly obvious that there is a La Niña signature of lower-than-normal chlorophyll concentrations in the spring and summer of 1998, which is also what we expected.

The milder El Niño/La Niña event in 2002-2003 is not as clearly defined. It appears that the winter season had slightly lower chlorophyll concentrations, particularly in the region between 5 and 6 degrees North. But between 3 and 4 degrees North, chlorophyll concentrations appear higher than in any other year! Could both of these observations be part of the El Niño/La Niña event?

Possibly. But there is one other interesting aspect of this event, which is well-illustrated by a time-plot for just the area between 6 and 5.4 degrees North:

In this plot, the reduction in chlorophyll concentration that occurred in 1997-1998 is very obvious, and the lower chlorophyll concentrations in winter 2001-2002 are also apparent. There is also a very interesting small peak in chlorophyll concentrations in the autumn of 2001. It is also possible to see this in the Hovmoller plots, but it is more obvious in this time plot. It is possible that this "spike" is an early indicator of the mild El Niño about to occur -- but it could also be something unusual that isn't related to El Niño.

7. Statement of Results

Our analysis indicates that there was an approximately 50% reduction in chlorophyll concentration during the winter period in the Gulf of Panama during the 1997-1998 El Niño. This season was followed by a noticeable La Niña period where chlorophyll concentrations were lower than average. During the milder El Niño/La Niña event of 2002-2003, winter chlorophyll concentrations appear slightly reduced over the whole region, with some subregions approaching a 50% reduction compared to normal years. Low chlorophyll concentrations characteristic of La Niña appear in summer and autumn 2002, and also in 2003. There is an anomalous increase in chlorophyll concentration in late summer 2002 occurring near to the Panamanian isthmus which may be an El Niño event precursor.

8. Discussion of Results

So now we have a clear statement of what has been observed in our data analysis. The next step is to examine explanations for these observations. We will discuss two possible explanations, but scientists should always consider a variety of explanations that could account for their observations.

The reduction in chlorophyll concentration occurring during the El Niño event could be caused by a change in the weather patterns that cause the strong winds which in turn cause mixing and nutrient availability in the Gulf of Panama. El Niño events are known to produce more storms on the U.S. West Coast and more rain in the U.S. Southwest, so this meteorological shift might also change the timing and intensity of the high pressure systems that enter the Caribbean Sea, which induce the strong winds through the passes in the Central American mountains. If these winds were not as strong or as persistent as in normal years, the mixing would be decreased, and this would also decrease the concentration of nutrients, leading to a reduction in phytoplankton productivity, which is visible as chlorophyll concentration.

Another explanation, which based on the oceanographic characteristics of El Niño events may be the most likely, is the increase in the depth of the thermocline (the depth boundary between cold subsurface water and warm surface water in the oceans). El Niño events move a large amount of warm water from the western Pacific Ocean to the eastern Pacific Ocean, which has the effect of "deepening" the thermocline, and reducing the intensity of upwelling over large areas of the Pacific. If the thermocline depth was increased in the Gulf of Panama, even if the winds were just as strong as in non-El Niño years, the effectiveness of the mixing for bringing nutrients to the surface would be lessened. The cold subsurface waters with higher nutrient concentrations would lie deeper, requiring more wind energy to mix the water column and bring them to the surface. Because the increased thermocline depth is the primary cause of reduced productivity in the Pacific Ocean during an El Niño event, it seems plausible that this is also the cause of reduced productivity in the Gulf of Panama during an El Niño event.

9. Statement of Conclusions

Now we provide a succinct statement of our conclusions.

• Chlorophyll concentrations were observed to be lower in the Gulf of Panama during the winter productive season for El Niño events occuring in 1997-1998 and 2002-2003.
• During the strong 1997-1998 El Niño, concentrations were lower by approximately 50%. During the 2002-2003 El Niño, concentrations were slightly reduced, and in subregions the reduction in concentration approached 50%. However, in other subregions concentrations were slightly higher for the 2002-2003 El Niño.
• Slightly elevated chlorophyll concentrations in the Gulf of Panama during summer and autumn 2001 may have been a precursory indication of the El Niño event.
• The likeliest explanation of the reduced chlorophyll concentrations is an increased depth of the thermocline, reducing the effectiveness of wind mixing for bringing high nutrient subsurface water to the surface. An alternative explanation is a shift in the pattern and intensity of high pressure systems in the Caribbean Sea that create high wind conditions in the passes of the Central American cordillera, which induce wind mixing of the waters in the Gulf of Panama.

10. Questions for Further Investigation

There is always more that can be done. While GIOVANNI provides a rapid way to examine chlorophyll concentration data, it doesn't allow more sophisticated numerical analyses. It would be quite interesting to integrate the chlorophyll concentration data (i.e., to sum up the chlorophyll concentrations over the area and over time) to provide a better quantitative estimate of how much chlorophyll concentrations were reduced in the Gulf of Panama during the El Niño events. This would help to confirm that our visual impression for the 2002-2003 event of slightly reduced productivity is correct.

To determine if the cause of the reduced chlorophyll concentrations is due to the increased depth of the thermocline or to a change in the wind mixing intensity, it would be necessary to obtain oceanographic data that measure the temperature of the water at various depths. Either ships that investigated the region or moored oceanographic buoys might provide these data. To examine the wind patterns, meteorological data for the Gulf of Panama and the Panama Canal (the pass that the winds blow through) could be examined for a comparison of a normal winter to an El Niño winter. This additional data would add confidence to our explanations of the observations we have made.

Finally, it would also be interesting to see if there are any data indicating a "spike" in chlorophyll concentrations before the 1997-1998 El Niño. SeaWiFS began collecting data in mid-September 1997, so data from summer 1997 would be required. Unfortunately, the Japanese satellite carrying the Ocean Color and Temperature Scanner (OCTS) instrument, which also collected ocean color data, failed in June 1997, so there is no satellite data for this important period. There might be data from local research or fishing vessels, however, that could provide information to fill this difficult gap in satellite data coverage. Also, because the chlorophyll concentration "spike" might also be visible in sea surface temperature (SST) data, it might be possible to use this type of data to figure out what was happening before September 1997.