Climate change is impacting the production of agricultural goods across the planet, and Vietnam is no exception. Vietnam's agriculture sector is estimated to contribute just under 20 percent to Gross Domestic Product (GDP) , a common measure of national economic output, and employs around half of the country's workforce. As a result, Vietnam's economic performance is tied, in part, to the success of its agriculture sector. However, climate change is causing more extreme weather events to happen, and the rise in extreme weather events can have disastrous consequences for agriculture. Farmers' crops may be destroyed and consumers may face higher food prices due to reduced supply. Furthermore, the negative impacts from adverse weather can inhibit agricultural production in the following year, as the agricultural land tries to recover from these adverse shocks.
These extreme weather events also cause problems for local, regional, national, and global economic development. As agriculture plays a significant role the economy in many regions, such as Vietnam, a decline in agricultural production would be expected to negatively impact economic growth, as measured by the GDP. This article presents a brief study highlighting the relationship between changes in Vietnam's weather over the past 10 years and the country's GDP over that same period, by examining aWhere's robust and complete weather database. This study is conducted in R. We have adopted a portion of the U.S. Climate Extremes Index, and use this framework to create an index of extreme weather, using aWhere's data. In the course of our survey, we found a strong relationship between this index and the rate of economic growth in Vietnam over the past 5 years. GDP figures for Vietnam are taken from the Government of Vietnam .
Using aWhere's global daily weather data, we calculated the historical Long Term Normal (LTN) weather across Vietnam, every day for the past decade. This was used to determine if the weather on a given day was significantly different from the LTN. The index defines extreme conditions as those falling in the upper or lower tenth percentile of the historical record. The index was then calculated by taking the arithmetic mean of the following 4 indicators from the U.S. Climate Extremes Index:
- The percentage of Vietnam with maximum temperatures much below or above normal.
- The percentage of Vietnam with minimum temperatures much below or above normal.
- The percentage of Vietnam with a much greater than normal number of days either with and without precipitation.
- Twice the value of the percentage of Vietnam with a much greater than normal proportion of total precipitation derived from extreme precipitation events. Extreme precipitation events are defined as being in the highest tenth percentile of precipitation amounts.
Each indicator has an expected value of 20 percent. For the purpose of this study, we will focus upon the temperature component of this index.
Measures of Gross Domestic Product (GDP)
Figure 1: The above plot shows the inflation adjusted value of both overall GDP and agricultural GDP in Vietnam over the years.
Figure 2: The above plot shows how closely the overall GDP growth rate and the agricultural GDP growth rate move together in Vietnam.
Figure 3: The above plot shows the agricultural GDP growth rate in Vietnam in recent years.
Looking at the time series plot of both Vietnam’s overall GDP and agicultural GDP, one can see that the two series are highly cointegrated. This cointegration supports the hypothesis that the rise in the rate of extreme weather events due to climate change and their subsequent negative impact on the countries agricultural economy will have an impact on the overall rate of economic growth in Vietnam.
According to the trace test statistic for r = 0 and r <= 1 at the 1%, 5% and 10% confidence levels, both of the test statistics exceeded the 1% confidence level, leading us to reject the null hypothesis of no cointegration. We, therefore, conclude that there is cointegration between the two time series.
Constructing a Simple Index
Next we will use aWhere’s global and robust weather data to construct the temperature portion of the U.S. Climate Extremes Index.
To construct our index, we compare Vietnam’s annual temperature data to the long-term normal (LTN) data. We did this for each of aWhere’s nearly 4,000 virtual weather stations across Vietnam. This procedure can be done using the sample code included below. In this code, we calculate the range of normal maximum and minimum temperatures (10 - 90%) for each location in the country for every day of the year. Next, we compare those derived daily values to the maximum and minimum temperatures observed in the same locations, to determine if any values were abnormal enough to be labeled “extreme.” Notably, the index only uses weather data for days prior to the day being compared, so that the analysis is representative of what would have been measured had the analysis been run in real time.
The table below shows an example of the output from the above procedure for a single virtual weather station. For this location, during the period October 1st-9th, 2016, the maximum temperature was within the typical range, based on the LTN, on all days except October 7th. The maximum temperature on that day was 24.45C while the normal maximum temperature range for that day is 14.82C - 23.94C.
Having now repeated that same calculation for the other 3,955 virtual weather stations in Vietnam, we can produce histograms using data from all of the locations, aggregated by year, to take a closer look at different frequencies of abnormal temperature events on a year to year basis.
Figure 4: The above figure shows the share of locations across Vietnam that experienced abnormal or extreme weather related to min/max temperatures. All regions above 0.2 threshold are considered to be experiencing extreme conditions.
Figure 5: The above figure shows the number of locations across Vietnam that experienced a specific abnormal temperature index value.
Most of the locations in Vietnam have experienced abnormal maximum temperatures approximately 30 percent of the time between 2011 - 2015 when compared to previous years. Abnormal temperatures are defined as falling outside of the 10th-90th percentile range of the LTN. For minimum temperature, a bimodal distribution was observed in the majority of years.
As stated above, the purpose of this study was to determine if there is a relationship between the rise of extreme weather events in Vietnam and changes in that country’s economic growth rate. To do this, we now aggregate our temperature-based index to the national level. While we would expect the effects of abnormal weather to be seen in the raw GDP numbers in the current year, we also expect the effects to be seen in the next year. This is because abnormal weather may damage the agro-ecosystem that agriculture relies on, due to impacts on soil, biotic life, and ecosystem dynmaics, and, therefore, will limit the potential of the coming growing seasons while the agro-ecosystem recovers.
Because of this expected pattern, we lag the GDP data with respect to our derived index for the remainder of the study.
Figure 6: The above plot shows that when there is more abnormal weather (the mean_index is higher) the agricultural GDP growth rate slows and when there is less abnormal weather (the mean_index is lower) the agricultural GDP growth rate increases.
From the above plot, we can see that when the stress index increases the rate of agricultural GDP growth slows and vice versa. This plot suggests a negative correlation of -0.99 between the derived stress index and agricultural GDP growth rates when looking at 4 years of data. However, testing the robustness of this correlation would require further analysis with additional data points. If the suggested correlation is accurate, it suggests that extreme weather events would impact total GDP rates at the national level.
This type of analysis, using derived stress indices computed from aWhere’s weather data, can be applied to understand the economic impacts of extreme weather events at a national, regional, and even local level. Because aWhere’s data is global in coverage and complete over 15+ years, studies across both time and space can be conducted that seek to better understand how changes in the weather are driving changes elsewhere on the planet.