Rainfall Distribution Trend Analysis of the Philippine National

Capital Region (2013-2016)

Miguel Aaron M. Bobadilla, Ryan Gabriel A. Eugenio and Maria Teresa R. Pulido

Department of Physics, Mapúa University, Intramuros, Manila City, 1002, Philippines

Keywords: Natural Science as a Service (NSaaS), Weather Forecasting, Decision as a Service, Big Data as a Service

(BDaaS), Big Data Algorithm, Trend Analysis, Rainfall Distribution, Mann-Kendall Test.

Abstract: The Philippine archipelago is a tropical country that experiences only two major seasons annually: wet

(June-November) and dry (December-May). Due to these conditions, the country is bound to experience

significant amounts of rainfall, followed by drought. Hence, studying long-term rainfall trends is highly

beneficial for the country’s livelihood and safety. In this work, we studied the rainfall distribution in the

National Capital Region covering the period of 2013 to 2016, and analysed the data using the Mann-Kendall

Test and the Bootstrap procedure. Using a monthly scale, we found a negative trend, signifying a decrease

in rainfall amount over the four years of data. Interestingly, we found a positive trend using a yearly scale,

showing an increase of rainfall overall. Therefore it is quite risky to generalize a certain region's rainfall

condition just by looking at it annually, but must consider as well its seasonal and monthly phenomena for a

more detailed analysis. We note also that the area being studied was considerably large and the rainfall data

varied with the location of the weather station where it was obtained. This work demonstrates the potential

of using Big Data and the Internet of Things to measure and predict weather trends using various sensors

and processors.

1 INTRODUCTION

The Philippines in general experiences most often

record breaking typhoons as years go by which

brings heavy rainfalls that consequently destroys a

lot of properties mainly its agricultural resources

which lead to production losses (Lansigan, 2013;

Cinco, et al., 2016).

Because of this, rainfall activity poses a great

point of attention in making a good preparation of

what's to come to our distant future. This kind of

study has actually been constantly the point of

change assumption in a span of 30-65 years in the

making particularly about climate change itself

(Cinco, 2018). The climate trends accounted in the

said studies also made use of the Mann-Kendall non-

parametric test since it's basically used for

identifying trends in time series data.

To further supplement this concept, we made use

of the Bootstrap procedure used to study rainfall

trends in Sicily (Cannarozzo, et al, 2006). The trends

found within a span of several years will be

categorized in a temporal manner (yearly, monthly),

eliminating the autocorrelation of time series data

that defeats an essential application of Mann-

Kendall Test.

The Philippine Atmospheric, Geophysical,

Astronomical Services Administration (PAGASA) is

the national agency mandated to “provide protection

against natural calamities and utilize scientific

knowledge as an effective instrument to insure the

safety, well-being and economic security of all the

people, and for the promotion of national progress”

(Presidential Decree No. 1149, 1977). This agency

has gathered, among many others, rainfall data over

several years and has made it public via the Freedom

of Information website (https://www.foi.gov.ph/).

We chose to collect data regarding the Philippine

National Capital Region (NCR) due to its political

and economic significance.

2 METHODOLOGY

2.1 Dataset

We used the rainfall data collected from three

PAGASA weather stations located in the NCR

370

Bobadilla, M., Eugenio, R. and Pulido, M.

Rainfall Distribution Trend Analysis of the Philippine National Capital Region (2013-2016).

DOI: 10.5220/0007750303700375

In Proceedings of the 4th International Conference on Internet of Things, Big Data and Security (IoTBDS 2019), pages 370-375

ISBN: 978-989-758-369-8

(Ninoy Aquino International Airport in Pasay City,

Science Garden in Quezon City, and Port Area in

Manila). These data were made available to the

public via the Freedom of Information website.

Figure 1: NCR Map showing 3 PAGASA weather

stations: Science Garden in Quezon City, Port Area in

Manila City, and NAIA in Pasay City.

The four year-long daily data acquired from

PAGASA are designated in the three following

weather stations as shown in Figure 1. From these

selected stations, only NAIA station have missing

data comprised of about 0.55% of its total and the

consequent changes from over all stations shown

were expressed in terms of its average monthly and

annual rainfall amount.

To further show the variability of data, modified

version of Oliver's Precipitation Concentration Index

(PCI) (Oliver, 1980) was used which analyzes

yearly, as well as monthly heterogeneity of rainfall

amounts. This index is described as:

100∙

∑













∑













(1)

where p

is the rainfall amount for the i

order of n

which is either equal to twelve if analyzed annually

and 29, 30, or 31 if the case is monthly analyzed.

The resulting PCI with values below 10 is said to be

uniform throughout the particular time scale, while

for 11 to 20 values would mean seasonality in

rainfall amounts, and values of 20 above

corresponds to substantial variability in the specific

time scale.

2.2 Trend Detection

The Mann-Kendall (MK) test is often used to detect

trends in a time series dataset with no apparent

pattern, such as rainfall. MK does not require the

data to be normally distributed, making it a non-

parametric or distribution-free test. In using the MK

test, we assume that there is no trend present, that

the observations over time are representative of the

true conditions at sampling times, and that the data

to be analyzed is unbiased and representative of the

underlying populations over time (Mann, 1945). Our

assumed Null Hypothesis (H

) is that there is no

existing trend from the rainfall data, while the

Alternative Hypothesis (H

) is that there is an actual

monotonic trend existing in the data. To either

accept or reject these hypotheses, we have



∑∑



















(2)

and specifically,

signT



T







1ifT



T



0

0ifT



T



0

1ifT



T



0

(3)

where T

and T

are the values in the times j and i

respectively, j > i. A positive value of S implies that

a majority of the differences between earlier and

later measurements are positive, suggestive of an

upward trend over time. Likewise, a negative value

for S implies that a majority of the differences

between earlier and later values are negative,

suggestive of a decreasing trend. A value near zero

indicates a roughly equal number of positive and

negative differences.

In conjunction with the MK test is the bootstrap

procedure or approach which counteracts the

correlation of time series data of rainfall amounts.

We extract randomly one year of a time series,

rearrange randomly the data for that particular year,

then compare the resulting statistic S to the original

unarranged data for the year.

Rainfall Distribution Trend Analysis of the Philippine National Capital Region (2013-2016)

371

Figure 2: Daily rainfall in NCR averaged on a monthly basis.

Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec

Rainfall (mm)

Average Rainfall - 2013

NAIA

Port Area

Science

Garden

Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec

Rainfall (mm)

Average Rainfall - 2014

Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec

Rainfall (mm)

Average Rainfall - 2015

Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec

Rainfall (mm)

Average Rainfall - 2016

IoTBDS 2019 - 4th International Conference on Internet of Things, Big Data and Security

372

3 RESULTS AND DISCUSSION

The daily rainfall was averaged over each month and

shown in Figure 2. Rainfall is seen to coincide with

the seasonal rainfall patterns prominent in the

months of June to November (wet season) and

December to May (hot season). Such trends are seen

more clearly in Figure 3, which averages the

monthly data from the three stations.

The trends in Figures 2 and 3 displayed

consistency with minor differences between years.

These differences are further expounded with the

PCI calculations and the MK test trends. In

particular, the yearly scaled trends in Figure 4

greatly vary when compared with the monthly scaled

trends in Figure 5. We note that the same method of

obtaining trends yields quite different results when

scaled over different time periods, which reminds us

to choose the region and scope of interest

appropriate to our study.

Figure 3: Daily rainfall in NCR averaged monthly for the three PAGASA stations.

Figure 4: Average yearly trend of rainfall in NCR for the three PAGASA stations.

Rainfall Distribution Trend Analysis of the Philippine National Capital Region (2013-2016)

373

Figure 5: Monthly scaled trends in rainfall data.

IoTBDS 2019 - 4th International Conference on Internet of Things, Big Data and Security

374

While the trends vary with the location of the

weather station, the yearly and monthly trends are

generally the same for all locations. For rainfall data,

monthly trends are more accurate than yearly trends,

since the former are more finely resolved and show

the variability of rainfall in the Philippines within a

year. The next step would be to use these trends to

predict data for the succeeding years, and determine

the presence of climate change with regard to

Philippine rainfall.

4 CONCLUSIONS

We can conclude that the MK Test is a very useful

tool for analyzing rainfall and other meteorological

data, which is optimized via Bootstrap resampling.

The analysis made over the NCR is fruitful since the

trends on temporal scale were successfully shown.

The variance in monthly and yearly trends remind us

of the inherent seasonal variability of rainfall within

a year for the Philippines. The specific location is

still an important factor in characterizing the

distribution for areas as large as the Philippine NCR.

We can extend this study using the remaining

PAGASA weather stations around the Philippines,

as well as other data sources. The emergence of Big

Data and the Internet of Things gives researchers

access to an unprecedented wealth of meteorological

data as well as the tools to measure possible trends

and predict future ones.

ACKNOWLEDGEMENTS

We thank the Philippine Atmospheric, Geophysical

and Astronomical Services Administration

(PAGASA) and the Freedom of Information website

for the rainfall data, the Mapúa University

Yuchengco Innovation Center for the resources in

preparing this manuscript, and our colleagues and

loved ones for their support. We also thank the

organizers of the IoTBDS 2019 Conference for

accepting this work and for the financial support.

REFERENCES

Cannarozzo, M., Noto, L.V. & Viola, F., 2006. Spatial

distribution of rainfall trends in Sicily (1921–2000).

Physics and Chemistry of the Earth, Parts A/B/C,

31(18), pp. 1201-1211.

Cinco, T. A., et al., 2016. Observed trends and impacts of

tropical cyclones in the Philippines. International

Journal of Climatology, 36(14), pp. 4638-4650.

Cinco, T. A., 2018. Observed Climate Change and

Projected Climate Change in the Philippines. [Online]

Available from: https://designingresilience.ph/wp-

content/uploads/day-1-observed-climate-trends-and-

projected-climate-change-in-the-philippines-pagasa-

15lzz9-file.pdf [Accessed 30 December 2018].

Lansigan, F. P., 2013. Climate risk for weather index-

based insurance for crop production in the

Philippines. [Online] Available from http://nap.psa.

gov.ph/ncs/12thncs/papers/INVITED/IPS-01%20Envi

ronment%20and%20Natural%20Resources%20Statisti

cs%20including%20Climate%20Change%20I/IPS-01

_1%20Climate%20Risk%20for%20Weather%20Index

-Based%20Insurance%20for%20Crop%20Production

%20in%20the%20Philippines.pdf [Accessed 30

December 2018].

Mann, H. B., 1945. Nonparametric tests against trend.

Econometrica: Journal of the Econometric Society,

pp.245-259.

Oliver, J. E., 1980. Monthly precipitation distribution: a

comparative index. The Professional Geographer

32(3), pp. 300–309.

Presidential Decree No. 1149 1977 (Establishing the

Philippine Atmospheric Geophysical and

Astronomical Services Administration), Malacanang,

Manila.

Rainfall Distribution Trend Analysis of the Philippine National Capital Region (2013-2016)

375