-
1. INTRODUCTION
Wischmeier and Smith (1978) described rainfall erosivity as an
interaction between kinetic energy of raindrops and soil surface.
This interaction may result to a greater or lower degree of
detachment and down-slope transport of soil particles according to
the amount of energy and intensity of rain by considering the same
type of soil, topographic conditions, soil cover and management.
The rainfall erosivity factor, R, is one of the important
parameters for the prediction of the erosive potential of raindrop
impact. It represents the climatic influence on water related soil
erosion (Hashim et. al, 2001) and can be used to quantify broad
scale, climate driven soil erosion potential.
The most commonly used method for predicting soil loss from
agriculture field due to water erosion is the universal soil loss
equation (USLE) (Wischmeier and Smith, 1978), and its successor,
the revised universal soil loss equation (RUSLE) (Renard et al.,
1997). To determine the soil erodibility parameters for the
USLE/RUSLE at a particular site for soil loss prediction, the
numerical values of rainfall erosivity factor, R, or EI30 values
are required. The original method to calculate the R factor needs
long pluviographical records at short intervals (Wischmeier and
Smith, 1978). Limited long term, continuous pluviograph data makes
it difficult to determine the R factor in many parts of the world
including Malaysia (Hashim et. al, 2001).
Besides, rainfall erosivity can be estimated using daily,
monthly and annual rainfall averages by other equations. Hashim and
Eusof (2001) developed a model to calculate the R factor for
Peninsular Malaysia using monthly rainfall data. They came up with
new values for the unknowns in the equation suggested by Yu and
Rosewell (1996). Morgan (1986) developed a relation between
rainfall erosivity and the annual precipitation. However, the use
of annual precipitation ignores the regional seasonality which in
some cases is necessary for two or more parallel analyses for
specific seasons. Hui (1999) used this approach in Malaysia to
produce soil erodibility nomograph. Loureiro and Coutinho (2001)
developed a procedure to estimate the EI30 values based on monthly
rainfall data and applied to the Algarve region in Portugal.
According to Alexandre (2003), many authors have found a good
relationship between the Fournier index (Fi) and annual values of
rainfall erosivity. The Fournier index is an equation relating to
the monthly values of precipitation for a month and the annual
values of precipitation. The Fournier index was developed and used
for seasonal rainfall erosivity (Cohen et al., 2005; Countinho and
Tomas, 1994).
The rainfall erosivity models can be used to show spatial
distribution after the calculation of erosivity values for each
station by interpolation using geographic information system (GIS).
For example, rainfall erosivity map for Brazil was prepared using
the
A. SHAMSHAD, W.M.A WAN HUSSIN, S.A. MOHD SANUSI, M.H. ISASchool
of Civil EngineeringUniversity of Science Malaysia14300 Nibong
Tebal, Pulau PinangMalaysia
KeyWords: RUSLE, Soil Loss, Rainfall Erosivity, Pluviographic
Data, Spatial Distribution.
SPATIAL AND TEMPORAL DISTRIBUTIONS OF RAINFALL AND EI30 VALUES
IN PULAU PINANG, PENINSULAR MALAYSIA
ABSTRACTSoil erosion is a serious problem for the agriculture
land use in Malaysia and many other parts of the world. In most of
the countries, Revised Universal Soil Loss Equation (RUSLE) is the
popular model for predicting the soil losses in runoff from
specific field areas in specific cropping and management system.
Rainfall erosivity factor (R) is one of the important parameters of
this model. As suggested in the past, the R factor or EI30 values
should be computed from the long term, continuous pluviographic
records. There is a lack of such kind of data in the world
including Malaysia. In contrast daily, monthly and other long
interval rainfall data are available at wider scale. In the present
study, an attempt has been made to develop rainfall erosivity
models using pluviographic data of six stations. The models can be
used to compute rainfall erosivity using monthly rainfall data.
Using the equations, the EI30 values were evaluated at other
stations using monthly rainfall data. The result showed that the
mean annual rainfall in the area varied from about 2200 mm to 3900
mm. The corresponding R-factor ranged from about 9,000 to14,000 MJ
mm ha-1 hr-1 year-1. For these tropical sites, both rainfall and
EI30 values are highly seasonal with two peaks, in general, in the
months of April and October.
18
PEER REVIEWED ARTICLE
-
Fournier models, linear model and exponential model developed by
different authors for various regions of Brazil (Silva, 2004). Qi
et al. (2000) generated rainfall erosivity map for the Republic of
Korea.
The present study emphasizes on the spatial and temporal
distributions of rainfall erosivity in the study area. One of the
main objectives of this study is to develop reliable methods for
estimating the EI30 values or the R factor for Malaysia and other
similar tropical regions of the world using the limited
pluviographic data.
2. STUDY AREA AND RAINFALL DATA
The present study has been carried out in the state of Pulau
Pinang, Peninsular Malaysia. It consists of an island and a part of
the mainland of the peninsular. Pulau Pinang is about 370 km from
the capital city Kuala Lumpur. Pulau Pinang is divided into 5
divisions namely Central Seberang Perai, North Seberang Perai,
South Seberang Perai, Northeast Seberang Perai and Southwest
Seberang Perai. It consists of an area of about 1031 sq km with
latitudes ranging from 58 to 535 and longitudes from 1008 to 10032.
The mainland covers about 738 sq km of the state. The elevation in
the area varies from 0 m to about 500 m. The main land use
categories in the area are oil palm, coconut and paddies.
In Malaysia the rainfall data is recorded and maintained by the
Malaysian Meteorological Services Department. The daily and monthly
rainfall data can be easily obtained; however for such type of
study rainfall, data at short intervals known as pluviographic data
is required. Pluviometric records at 15 minutes interval of 6
weather stations of Peninsular Malaysia is available for the
computation of EI30 values. The record lengths varies from 6 to 35
years, such as Simpang Ampat (1988-2004), Taliair Besar
(1970-2004), Kompleks Perai and Ibu Bekalan (1971-1975), and Klinik
Bukit Bendera and Kolam Bersih (1976-2004). In addition, the daily
rainfall data of 16 more stations were obtained.
3. METHODOLOGY
All of the continuous 15 minutes rainfall data of 6 stations
have been used to compute rainfall erosivity (EI) for each storm
event and a relationship between EI30 and the Fournier index (Cc)
is established for each station. Individual rainfall events of less
than 10 mm were excluded from the computation of EI. The annual
rainfall erosivity, R, is the sum of erosive storm EI30 values
occurring during a mean year. Individual rainfall events of less
than 10 mm which were separated from other events by more than 6
hours without rain were excluded from the computation of EI values
unless the depth of rainfall in 15 minutes exceeds 6 mm (Wischmeier
and Smith, 1978; Mannarerts and Gabriels, 2000).
Firstly, the kinetic energy of rainfall (Ej) is computed using
the formula derived by Onaga et.al (1998) for eastern Asia:
Ei = 9.81 + 10.6 log10 lj
where, Ei = kinetic energy of rainfall (J/m2) and lj = rainfall
intensity (mm/h).
The sum of the kinetic energy gives the kinetic energy of the
whole rainfall event (E). The rainfall erosivity (EI30) is the
kinetic energy of the whole rainfall event (E) multiplied by the
maximum 30-minute rainfall intensity (I30).
EI30 = E x I30 x 1/100
EI30 = rainfall erosivity (kJ/m2.mm/hr)
By summing the EI values for each storm, the total erosivity for
each month and year is computed.
4. RESULT AND DISCUSSION
4.1 Rainfall Distribution
To study the annual and monthly rainfall patterns, the time
series plots of total yearly rainfall, monthly mean and maximum
values of rainfall were plotted for all the stations. For examples,
such plots for Kolam Bukit Berapit and Simpang Ampat are presented
in Figures 1a and 2a. All the rainfall stations show similar
patterns of rainfall for the years 1996 to 2004. For most stations,
the rainfall is high in 1999 and 2003 but low in 2002. However, the
annual rainfall curve is oscillating about the mean, thereby
indicating that there is no significant trend in the data, and the
mean annual rainfall is almost the same. Furthermore, the mean
monthly rainfall values at all the stations fall in the range of
100 mm to 250 mm.
It is observed that the months of April and October are the
raining seasons in this area, which is not the case throughout
Malaysia. October brings the most rainfall to this region as
compared to other months. Overall, the temporal patterns of
rainfall distribution are the same for all stations. This is
because these stations are situated in the same area.
The probability graphs of monthly rainfall for each station are
studied to show the possibility of a certain amount of rainfall
occurring in an area according to the location of the station. The
graphs for Kolam Bukit Berapit and Simpang Ampat are plotted in
Figures 1b and 2b. The figures reveal that the rainfall between 100
mm to 200 mm per month has the highest probability of occurrence.
During the period of study, rainfalls of about 900 mm per month
also occur.
To study the mean monthly and annual spatial distributions of
the rainfall data, ArcView GIS was used to interpolate using the
Inverse Distance Weighted (IDW) interpolator method in which input
point has a local influence that diminishes with distance. It
weights the points closer to the processing cell greater than those
further away. The resulting maps of rainfall for the month of
January, February, June, October and
19
-
December and the annual rainfall were presented in Figure 3. The
results show that the rainfall at all the stations is highly
seasonal with two peaks in the months of April and October and
ranges from about 80 mm to 250 mm except in the months of April and
October.
The rainfall is observed to be highest at most of the stations
in October, varying from 320 mm to 660 mm followed by April in
between 150 mm and 300 mm. The annual rainfall is least (2130 mm)
at Ladang Malakof and highest (3850 mm) at Bukit Bendera.
20
Figure 1a. Total yearly, monthly mean and maximum values of
rainfall at Kolam Bukit Berapit
Figure 2a. Total yearly, monthly mean and maximum values of
rainfall at Simpang Ampat
Figure 2b. Probability distribution of monthly rainfall at
Simpang Ampat
Figure 1b. Probability distribution of monthly rainfall at Kolam
Bukit Berapit
Figure 3. Spatial distribution of monthly and annual
rainfall
-
Further, to estimate the monthly and the annual erosivity for
the Peninsular Malaysia or even for the whole country or other
similar tropical regions of the nearby country, a common equation
was derived in this study by combining the data of all the six
stations (Figure 4). The model can be represented by the equation 2
as given below:
EI30 = 227.0 Cc0.548, R =0.90 (2)
Besides the seasonal (monthly) values of rainfall erosivity, the
storm erosivity values are needed to study the non-point source
pollution based on a particular storm event. One example of such
study is the event based analysis in Agriculture Non-Point Source
Pollution model (AGNPS) (Bhuyan, 2002). Using all the EI30 values
based on erosive storms of all the six stations under study, a
model was developed for rainfall erosivity of a single storm
(EIstorm) as given by the following equation:
EIstorm = 0.537 P1.736, R = 0.77 (3)
Similar types of models were developed by several other authors
to compute storm erosivity based on a single storm (Bhuyan, 2002;
Mannaerts and Gabriels, 2000).
21
4.2 Development of Erosivity Models
The monthly EI30 values were computed for each of the 6 stations
(Table 1) using the pluviographic data. Fournier Index (Fi) for
each month was determined using the following equation (Cohen et
al., 2005):
(1)
where, Mi is the average monthly precipitation depth (mm) and P
is the average annual precipitation (mm). Regression analysis was
carried out between EI30 values and the Fournier Index. Three
different types of equations were obtained for every station to
compare and give us the best equation at the end. The three
different types of derived equations are the linear, logarithmic
and power. After all equations are derived, comparisons between all
equations for all the stations were made. From the results obtained
using the Microsoft Excel (and checked by using SPSS software), the
equations with the highest correlation are the equations of power.
A comparison carried out based on the error analysis also produced
the best result with power equation at almost all the stations.
These equations for EI30 values with Fournier index of all the six
stations have been presented in Table 2 along with their
coefficients of correlation (R). The values of R are mostly above
0.90 which are very good values. This showed that the accuracy of
the unknowns found is high and the difference between the observed
and the computed value is little. However, this is not true for one
station which is located in Kompleks Perai whereby the R is 0.88.
Though this value is low, it is still acceptable when compared to
the values calculated by Banasik and Gorski (1998) in their study
for East and Central Poland.
Station Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Annual
(R)
Simpang Empat 592.0 572.5 374.5 717.5 402.9 236.2 532.7 676.9
958.5 1122.6 1229.1 544.8 7960.2Taliair Besar 197.2 787.6 1084.4
1746.5 1517.4 1032.6 1326.5 1332.0 2102.9 2569.6 1535.5 579.9
15812.1
Kompleks Perai 199.3 631.9 448.7 1283.3 1318.4 2053.5 695.6
825.0 1263.7 2645.5 1349.3 644.0 13358.2Klinik Bukit Bendera 386.8
464.3 1219.2 1830.4 1392.7 1251.7 1683.6 1926.8 2248.8 2492.9
1543.7 575.7 17016.6
Kolam Barsih 311.4 368.7 830.8 1051.4 1258.3 877.1 1101.5 1151.9
1864.4 1538.4 1386.4 462.6 12202.9
Ibu Bekalan 229.7 581.6 861.1 1826.2 1737.3 404.9 1038.2 871.5
2389.3 2235.8 1452.5 1055.1 14683.2
Table 1 EI30 and R (MJ mm ha-1hr-1) values computed using
pluviographic data at 15 minutes interval
Table 2 Fournier models developed using pluviographic data
Station Fournier Model R
Simpang Ampat EI30 = 119.0 Cc0.635 0.92Taliair Besar EI30 =
238.3 Cc0.579 0.98Kompleks Perai EI30 = 265.6 Cc0.535 0.88Klinik
Bukit Bendera EI30 = 311.4 Cc0.487 0.98Kolam Bersih EI30 = 280.9
Cc0.458 0.99Ibu Bekalan EI30 = 204.6 Cc0.658 0.95
Figure 4. Relationship between erosive storm EI30 values and
storm rainfall (P)
P
MF ii
2
=
-
5. CONCLUSION
Rainfall erosivity factor, R is the average annual summation of
EI30 values in a normal year's rain or in other words, it is the
quantitative expression of the erosivity of local average annual
precipitation and runoff. This study addresses the problem of
predicting the rainfall erosivity factor (R) of the Revised
Universal Soil Loss Equation (RUSLE) with limited rainfall
data.
In this study, equations were derived using two different ways
for EI30 values while using the monthly and storm rainfall data.
One of the ways related to the rainfall erosivity factor is through
the Fournier Index. The best equations derived are the ones with
power relationship. The values of coefficient of correlation (R)
for all the stations using these power equations range from 0.88 to
0.99. A general equation was then derived in this study which can
be used to calculate the monthly values of rainfall erosivity
factor even if the available data is monthly and not pluviographic
data. Another way of deriving the equations is to get a
relationship between storm rainfall erosivity with storm rainfall.
The value of coefficient of correlation (R) for this equation is
only 0.77 which is acceptable for such type of studies. The
rainfall erosivity (R) map derived using the model developed shows
that the monthly rainfall erosivity EI30 values are particularly
high during the months of April and October for the overall results
of all stations which are 1270 MJ mm ha-1h-1 and 2100 MJ mm
ha-1h-1, respectively. This is in accordance with the rainfall in
these locations. A higher value of rainfall produces a higher value
of rainfall erosivity factor. The soil erosivity maps developed in
the study are useful soil for conservationists, agronomists and
civil engineers. It is recommended that in further research,
rainfall erosivity maps should be developed for the whole of
Malaysia using pluviographic data covering a wider spectrum of the
region.
ACKNOWLEDGEMENT
The Authors wish to thank the Ministry of Science, Technology
and the Environment, Malaysia (MOSTE) and Universiti Sains Malaysia
(USM) for the financial support under the IRPA grant for the
project GIS Based Watershed Management System for Non-Point Source
(NPS) Pollution Modelling to carry out this study.
REFERENCES
Banasik Kazimierz and Gorski Dariusz (1998). Estimating the
Rainfall Erosivity for East and Central Poland. Warsaw Agricultural
University, Department of Water Engineering, Sedimentation
Laboratory, Warsaw, Poland.
Bhuyan, S.J., Prasanta, K.K., Janssen, K.A., Barnes, P.L., 2002.
Soil loss predictions with three erosion simulation models.
Environmental Modelling & software 17, 137-146.
22
Using the model developed above (i.e. Equation 2), the monthly
EI30 values were calculated at 16 other stations using the monthly
rainfall data. The spatial and temporal variabilities of rainfall
erosivity were studied in detail in the study area. It was observed
that the rainfall erosivity is highly variable from one month to
the other. This is mainly due to the climatic situation during a
certain time. Malaysia is known for their hot and rainy seasons and
these vary according to months. For this study area, the months of
April and October show more rainfall than the other months. Hence,
the values of the rainfall erosivity EI30 during these months are
higher than the rest of the months. The EI30 values during the
months of April and October for the overall results of all stations
are 1270 MJ mm ha-1hr-1 and 2100 MJ mm ha-1h-1, respectively.
Whereas, the months of January, February and December show the
lower values of rainfall erosivity which are about 575 MJ mm
ha-1hr-1. This shows that for these months, the total rainfall is
much lesser. As mentioned earlier, EI30 is the amount of energy
applied to the land surface due to rainfall energy from the falling
rain drops. It can be said that if there is more rainfall in a
certain area, the probability of erosivity is higher.
Figure 5 shows the annual rainfall erosivity, R, distribution in
the study area. The erosivity values vary geographically. The
annual rainfall erosivity calculated for all stations in the study
area range from about 9,000 to 14,000 MJ mm ha-1hr-1. The station
with the highest total annual rainfall erosivity is Bukit Bendera
with R-factor of 14,000 MJ mm ha-1hr-1 year-1. This means that this
area has a higher possibility to experience higher soil erosion
compared to the other stations. Whereas the stations at Simpang
Ampat and Ladang Malakof show the least annual rainfall erosivity
factor which is 9,000 MJ mm ha-1hr-1 year-1. These stations have
the least annual rainfall. Rainfall erosivity map created by GIS is
an important map to be referred to by soil conservationists,
landuse planners and civil engineers in studying the distribution
of erosivity of a particular area.
Figure 5. Annual rainfall erosivity (R) map
Contours of Rainfall Erosivity (R)Rainfall Erosivity (R)
9000 - 1000010000 - 1100011000 - 1200012000 - 1300013000 -
14000No Data
-
Morgan, R.P.C., 1986. Soil Erosion and Conservation. Longman
Group, Essex, UK
Qi, H., Gantzer, C.J., Jung, P.K. Lee, B.L. (2000). Rainfall
erosivity in the Republic of Korea. J. Soil Water Conservatin 55,
115-120.
Renard, K.G., Foster, G.R., Weesies, G.A., McCool, D.K., Yoder,
D.C. (1997). Predicting soil loss by water: A guide to conservation
planning with the revised soil loss equation (RSULE). Handbook, Vol
703. US Department of Agriculture, Washington, DC, USA.
Silva da, A.M. (2004). Rainfall erosivity map of Brazil. CATENA
57, 251-259.
Wischmeier, W.H., Smith, D.D. (1978). Predicting Rainfall
Erosion Losses. Agric. Hbk 537.U.S.D.A. Sci. and Educ. Admin.,
Washington, DC.
Yu, B., Hashim, G.M. and Eusof, Z. (2001) Estimating R-factor
with limited rainfall data: A case study from Peninsular Malaysia.
Journal of Soil and Water Conservation 50 (2), 101-105.
23
Cohen M.J., Shepherd, K.D., Walsh, M.G., 2005. Empirical
formulation of the universal soil loss equation for erosion risk
assessment in a tropical watershed. Geoderma 124, 235-252.
Countinho, M.A. Tomas, P.P., 1994. Comparison of Fournier with
Wischmeier rainfall erosivity indices. In: Rickson, R.J. (Ed.).
Conservation Soil Resources, European Perspectives. CAB
International, Wallingford
Hashim B. Yu. G.M and Eusof Z. (2001). Estimating the r-factor
with limited rainfall data: A case study from Peninsular Malaysia.
Journal of Soil and Water Conservation, Vol 56, No.2.
Hui, T.K., 1999. Production of Malaysian soil erodibility
nomograph in relation to soil erosion issues. Perputakan Negara
Malaysia.
Loureiro, N.D.S., Coutinho, M.D.A., 2001. A new procedure to
estimate the RUSLE EI30 index, based on monthly rainfall data and
applied to the Algarve region, Portugal. Journal of Hydrology 250,
12-18
Mannaerts, C.M., Gabriels, D., 2000. Rainfall erosivity in Cape
Verde. Soil & Tillage Research 55, 207-212.