April Showers Bring May Flowers, But What About Bone Dry Marches?

March in the Kerinci Valley has been particularly hot and dry this year.  Though I've only been here about 6 months, a good friend of man that's been working here for approximately two decades remarked to me that March is usually one of the wettest months of the year, and though there are sometimes dry years, this year has been particularly dry.  This is just latest manifestation of what farmers and conservationists alike see as increasingly erratic weather patterns with some folks suggesting a strong relationship between high levels of deforestation in the wider region and decreased rainfall in the valley.  The day after my friend's comment I happened to see an article in Kompas, the "paper of record" for Indonesia, attributing some strange weather in the archipelago to a phenomenon known as the Madden-Julian Oscillation (MJO).  I'd learned a bit about the MJO in some of the climatology classes I took at UH (1), so my curiosity was piqued.  I'd also learned that when looking at climate patterns and trends, it's important not to jump to sudden conclusions, because what appears to be an anomalous blip in the short term could be a normal part of long-term variability.  Was the abnormally low rainfall for July caused by regional deforestation, or was it part of a larger, regular climate pattern?

The Madden-Julian Oscillation

Graphic from here.

The Madden-Julian oscillation was discovered by a couple of climatologists (guess their names) in 1971.  They noted a 40-50 oscillation in tropical winds, which previously vexed climatologists as too variable to predict.  In many parts of the tropics there are two seasons: wet and dry.  Changes in the prevailing winds, known as monsoons, bring air masses from different directions.  For example, on the Indian subcontinent, the rainy season happens when moisture-laden air blows in from the Indian Ocean.  The dry season happens when the winds shift and air comes instead from the north.  Indonesia experiences this pattern as well.  However, during the wet and dry seasons there is intra-seasonal variation in rainfall.  In other words, sometimes there is a period of dryness in the wet season and rain during the dry season.  Before Madden and Julian made their discovery no one could really explain why this happens.  The climatologists discovered that these fluctuations are driven by pressure waves originating in the Indian Ocean, and that these waves move.  The semi-regularity with which these waves emerge is related to the intra-seasonal variations in rainfall.

Graphic from here.

In the 1980s a lot more research was done on this phenomenon, and it came to be known that this intra-annual fluctuation helps to explain weather patterns in the tropics.  It involves changes in wind, sea surface temperature, cloudiness, and rainfall.  Now climatologists say that the pattern operates on a cycle of 30 to 90 days, meaning that a new wave in the oscillation emerges every 1-3 months.  The MJO begins in the Indian Ocean and moves to the east at about 5 meters per second.  Generally it is associated with the Indian Ocean and the Western Pacific region, but it affects weather patterns all over the planet.  In simple terms, the MJO can cause the weather in some places to alternate between wet and dry a few times over the course of a season.  I wanted to see if the MJO has anything to do with the strange weather patterns in the Kerinci valley, so I decided to initiate a little side project in addition to my dissertation: an attempt to analyze the rainfall patterns and compare them to the MJO record.  You know, just for funsies.

To begin my analysis I needed to get 30 years worth of monthly rainfall data to get an idea of cyclical and long term climatological trends for the valley.  This is the basic (but not only) difference between climatology and meteorology; whereas meteorologists focus on day-to-day atmospheric changes, climatologists are more interested in the bigger picture.  A climatologists can't tell you if it's likely to rain this afternoon, but he/she can tell you whether it's likely to be dry or wet in July 25 years from now.  For climatologists the accepted time period for determining general patterns is 30 years.

Getting the Data

Automatic rain gauge at the airport.  They
also have a manual gauge.  

My first stop was the regional statistics agency (Badan Pusat Statistik, BPS) in Sungai Penuh.  The BPS is the government agency responsible for tracking, compiling, and publishing data ranging from demographic and economic indicators to weather information.  Every year (theoretically) the BPS publishes a fact-book about its district, and these are generally kept in the library at the BPS office.  If you are a foreign researcher you need special permission to access the library; fortunately I had already established a relationship with BPS and I know the guys that work there.  Some districts make their data available online, but Kerinci hasn't yet reached this level of sophistication, so you have to personally go to the office. 

The records at BPS begin in 1985, and unfortunately there is a significant amount of missing data.  Since the rainfall data is collected at Kerinci's "airport", I decided I'd go right to the source.  I rode my motorcycle down to Depati Parbo Airport, which sits in the middle of a rice field and is reached by following a bumpy hardened gravel road from Sungai Penuh to the village of Hiang.  The Meteorology, Climatology, and Geophysics agency of the central government maintains a small station there mainly to collect data for the airport and the farmers of the region.  After receiving a tour of the installation, I asked for the rainfall data over coffee and other refreshments.  My new friend, the head of the station, was happy to oblige.  He even invited me over to his house for coffee, an offer I'll take advantage of next time I'm in his neighborhood.

Some Graphs and Stuff

The first thing I looked at was the annual rainfall record, which I've illustrated in the line graph below.  The average rainfall over the period for the valley was 1311mm, but as you can see there is a bit of variability.  In general it appears that there could be a downward trend, but without further analysis (which I haven't done yet) it is difficult to tell whether this is significant or not.  Also, I am reluctant to draw any conclusions based on 27 years (with two missing years), because if you go further back you may find that this apparent trend is part of some longer-term cycle, or is within normal long-term variability.  What I did notice, though, is that there doesn't seem to be any clear direct relationship between La Nina/El Nino years and annual rainfall.  Some La Nina years have lots of rain, some have below average rain, likewise with El Nino.  This doesn't mean that there is no El Nino/La Nina effect here because there could be some sort of combined influence with El Nino/La Nina working in tandem with some other pattern (like the MJO).    

I also noticed that this March has indeed been unusually dry, but it's not the driest March on record.  In addition, I also noticed that whenever there is an abnormally dry March, April tends to be wet.  So although everyone, including the region's farmers, are worried about the lack of rain, I think that the dry spell will end pretty soon.  The next step in the project will be to figure out a good, rigorous way to compare my rainfall data with MJO intensity and to see if there is any sort of connection there.  In the picture below you can see the intensity of the MJO spanning a period of more than four decades.  

Chart from here.  

We'll see how my little investigation goes.  It could prove to be beyond my capabilities.  Another possibility is that my wonderful academic adviser sees this post and sends me and angry email admonishing me to get back to work.  

Notes

(1)  Climatology is one of the major strengths of the Department of Geography at the University of Hawai'i.

References and for Further Reading

For an excellent source on the MJO check out the Madden-Julian Conversation blog here.