ArticlePDF Available

Dynamic downscaling and bias correction of rainfall in the Pampanga River Basin, Philippines, for investigating flood risk changes due to global warming

Authors:

Abstract and Figures

The Pampanga River Basin in Philippines suffers from floods due to typhoons or monsoonal rainfall every year. Assessment of changes in flood risk due to global warming is, therefore, an important issue for this flood-vulnerable basin. We studied possible changes in rainfall features under present (1979–2003) and future (2075–2099) climates using dynamic downscaling of MRI-AGCM experiments. The GCM projections were downscaled into a finer resolution of 5 km for hydrological simulation (catchment size of 10,500 km2). The downscaled rainfall overestimated the number of weak rainfall events, which subsequently resulted in overestimation of monthly rainfall. Bias correction was carried out for downscaled rainfall with reference to raingauge rainfall to perform cumulative distribution mapping. The simulation results found that monthly rainfall would change slightly between present and future climate conditions, with extreme rainfall very likely to increase in the future. The annual maximum 48 h rainfall with a 50-year return period may increase from 320 mm under the present climate to 470 mm (MRI-AGCM 3.2S) or 530 mm (MRI-AGCM 3.2H) under the RCP8.5 future climate scenario. This increase in extreme rainfall in the future would have a significant impact on this vulnerable river basin.
Content may be subject to copyright.
—106—
Hydrological Research Letters 10(3), 106–112 (2016)
Published online in J-STAGE (www.jstage.jst.go.jp/browse/hrl). DOI: 10.3178/hrl.10.106
Dynamic downscaling and bias correction of rainfall in the Pampanga River
Basin, Philippines, for investigating ood risk changes due to global warming
Tomoki Ushiyama, Akira Hasegawa, Mamoru Miyamoto and Yoichi Iwami
International Centre for Water Hazard and Risk Management, Public Works Research Institute, Japan
Abstract:
The Pampanga River Basin in Philippines suffers from
oods due to typhoons or monsoonal rainfall every year.
Assessment of changes in ood risk due to global warming
is, therefore, an important issue for this ood-vulnerable
basin. We studied possible changes in rainfall features under
present (1979–2003) and future (2075–2099) climates using
dynamic downscaling of MRI-AGCM experiments. The
GCM projections were downscaled into a ner resolution of
5 km for hydrological simulation (catchment size of
10,500 km2). The downscaled rainfall overestimated the
number of weak rainfall events, which subsequently resulted
in overestimation of monthly rainfall. Bias correction was
carried out for downscaled rainfall with reference to rain-
gauge rainfall to perform cumulative distribution mapping.
The simulation results found that monthly rainfall would
change slightly between present and future climate condi-
tions, with extreme rainfall very likely to increase in the
future. The annual maximum 48 h rainfall with a 50-year
return period may increase from 320 mm under the present
climate to 470 mm (MRI-AGCM 3.2S) or 530 mm (MRI-
AGCM 3.2H) under the RCP8.5 future climate scenario.
This increase in extreme rainfall in the future would have a
signicant impact on this vulnerable river basin.
KEYWORDS dynamic downscaling; regional climate
model; global warming; bias correction;
Philippines
INTRODUCTION
The Pampanga River Basin in central Luzon Island, the
fourth largest basin in the Philippines, plays key roles in
socio-economic activities of the country; it is the most
important grain belt and provides Metropolitan Manila with
97% of water resource from the Angat dam in the basin. The
Pampanga River Basin, however, suffers from oods due to
typhoons or monsoonal rainfall at least once a year which
cause disruption and damage; about 4,000 houses were
totally destroyed in the worst year, and the cost of damage to
infrastructure and agriculture was up to US$ 43 and 280 mil-
lion in the 2000 and 2011 ood events, respectively (Shrestha
et al., 2014). The Intergovernmental Panel on Climate
Change 5th Assessment Report (IPCC, 2013) reports,
“Future increase in precipitation extremes related to the
monsoon is very likely in South America, Africa, East Asia,
South Asia, Southeast Asia and Australia”. Therefore, the
impact of global warming on ood risk change is an impor-
tant issue for this ood vulnerable basin, and assessment of
ood risk change is essential to mitigate social damage and
prepare measures to increase the resilience of local commu-
nities.
Jose and Cruz (1999) studied global warming effects on
the Angat water reservoir based on three General Circulation
Models (GCMs) and a hydrological simulation model. They
projected a decrease in water resources from rainfall due to
global warming, and also stressed the importance of global
warming on the socio-economy of the Philippines for the
rst time. More recent studies have investigated future
changes in precipitation and temperature by downscaling
GCM projections with different models: Chotamonsak et al.
(2011) used the Weather Research and Forecasting (WRF)
model, Wang et al. (2013) used the Consortium for Small-
scale Modelling-Climate Limited-area Modelling (COSMO-
CLM) in the framework of the Coordinated Regional
Climate Downscaling Experiment (CORDEX), and Cruz
et al. (2016) used the Non-Hydrostatic Regional Climate
Model (NHRCM). However, all these studies were con-
ducted with models having a downscale resolution of about
50 to 60 km and their discussion focused on monthly aver-
age precipitation. Because 50-km resolution models are only
capable of resolving rain systems of more than 200 km,
those models could not resolve extreme rainfall such as
typhoon rainbands.
To examine ood risk and changes due to global warming,
rainfall information with much ner horizontal and temporal
resolutions is desired. In this study, dynamic downscaling
into 5 km horizontal resolution was conducted for the rst
time in this area, since this ner downscaling has been found
effective to improve the reproduction of heavy rainfall
(Kanada et al., 2008). This level of resolution is also ne
enough for streamow and inundation simulation for this
river basin (catchment size of 10,500 km2). The boundary
conditions were provided by the Meteorological Research
Institute-Atmospheric General Circulation Model (MRI-
AGCM) version 3.2. We further performed bias correction
of the downscaled rainfall, because rainfall includes signi-
cant biases even after downscaling. The effect of global
warming is discussed using extreme rainfall cases.
DATA
We used data from raingauge observation for validation of
downscaled rainfall, and GCM projections for the boundary
Correspondence to: Tomoki Ushiyama, International Centre for Water
Hazard and Risk Management under the auspices of UNESCO, Public
Works Research Institute, 1-6, Minamihara, Tsukuba, Ibaraki 305-8516
Japan. E-mail: ushiyama55@pwri.go.jp
Received 1 September, 2016
Accepted 29 October, 2016
Published online 30 November, 2016
©2016, Japan Society of Hydrology and Water Resources.
DYNAMIC DOWNSCALING OF PAMPANGA BASIN
—107—
conditions of a regional climate model. We obtained hourly
raingauge data collected in the Pampanga River Basin for 22
years over the periods of 1980–1982, 1993, and 1995–2012,
including frequent periods with data missing. The number of
raingauges was 9 at the beginning and increased up to 17 at
the end.
The boundary conditions of dynamic downscaling were
given by MRI-AGCM 3.2S (super high resolution in 20 km)
and 3.2H (high resolution in 60 km) (Mizuta et al., 2012) for
present (1979–2003) and future (2075–2099) climate pro-
jections based on the RCP8.5 scenario. The model was
designed to have quite a high resolution compared with
other Coupled Model Intercomparison Project Phase 5
(CMIP5) GCMs, and brought the advantage of better repro-
duction of mean precipitation in the monsoon area (IPCC,
2013). Since MRI-AGCM is an atmospheric GCM, sea sur-
face temperature (SST) was given by monthly mean obser-
vations (present) or observations plus future changes derived
from the CMIP5. MRI-AGCM 3.2S was provided with four
different SST distributions: multi-model ensemble (MME)
and clusters 1, 2 and 3 (Kitoh and Endo, 2016). MRI-AGCM
3.2H was provided with only one future projection with
MME SST. The surface variables were given in original res-
olutions of 20 and 60 km. However, three dimensional vari-
ables (U; zonal wind, V; meridional wind, T; temperature, Q;
specic humidity, Z; geopotential height) were given in 1.25
degree resolution in 12 layers from 1000 to 100 hPa.
METHODS
Dynamic downscaling was performed using the Weather
Research and Forecasting (WRF) model ver. 3.4.1, which is
a next generation mesoscale numerical weather prediction
model developed by the National Center for Atmospheric
Research (NCAR) and several other research institutes in the
United States (Skamarock et al., 2005). The model domains
were set up with double nesting frames with 67 × 67 × 40
grids in both outer and inner frames at 15 or 5 km horizontal
grid intervals and stretching vertical grids (Figure 1). For
parameterization schemes, the WRF single moment 3-class
(WSM3) microphysics scheme, thermal diffusion surface
physics, the Mellor-Yamada-Nakanishi-Niino level 2.5 plan-
etary boundary layer scheme were adopted. The sea surface
temperature used for the integration of GCMs was given as
the lower boundary condition. To select a cumulus parame-
terization scheme, which is critical for production of precip-
itation, we tested three different types: the default Kain and
Fritsch scheme (Kain and Fritsch, 1993), the Kain and
Fritsch scheme tuned by the Japan Meteorological Agency
(JMA) (Narita, 2008), and the Grell 3D ensemble scheme
with a shallow convection option (Grell and Devenyi, 2002).
After comparing rainfall simulated using the three schemes
for a period of several years, we adopted the third scheme, as
it demonstrated the best performance, producing rainfall
having good agreement with extreme rainfall from rain-
gauge observations. The downscaling calculations were car-
ried out separately with respect to each year from January to
December. We did not include any additional spinup periods.
However, we assumed that the rst week was negligible,
since one-week spinup time was long enough and the begin-
ning of each year was in the driest season.
Figure 2 compares average rainfall from raingauge obser-
vations and WRF downscaling during four recent major
ood events in the Pampanga River Basin. In the WRF sim-
ulation, the boundary condition was provided by ERA-
interim reanalysis. The simulated rainfall variation in the
September 2009 and September 2011 cases agreed well with
that of the raingauge observations, although the rainfall
peaks during the September 2009 event were overestimated.
These two events were due to typhoon passages. In the June
2011 case, the simulated rainfall trend did not necessarily
match the observed one, but it did reproduce rainfall peaks
of similar magnitudes. In the August 2012 case, the simula-
tion showed a fairly good reproduction of rainfall that was
superimposed on diurnal variation. The latter two events
were caused by the southwesterly monsoon. The comparison
in Figure 2 shows that the WRF model is reliable in repro-
ducing typhoon rainfalls, and it was also capable of repre-
senting the basic characteristics of monsoonal rainfalls, sug-
gesting the validity of using the WRF model for our study on
extreme rainfall.
Figure 1. Model domain for the Pampanga River Basin. The left and right panels show the WRF model domain of the outer
and inner frames, respectively. Color shadings show topographic height (m)
T. USHIYAMA ET AL.
—108—
RESULTS
Figure 3 shows seasonal variation in rainfall and fre-
quency of appearance, before and after downscaling, and
after bias correction. Figure 3a shows GCM rainfall where
bias is corrected based on APHRODITE data (Yatagai et al.,
Figure 3. Seasonal variation in basin average rainfall (a)–(c), and frequency of appearance of daily rainfall at raingauge sites
(d)–(f). The rst column shows rainfall from GCM, the second column shows rainfall after downscaling, and the third column
shows rainfall after downscaling and bias corrected based on raingauge data. The GCM rainfall in (a) were bias corrected
based on APHRODITE rainfall. In (a)–(c), the black line indicates observations (APHRODITE in (a), and raingauge in (b) and
(c)), the blue line indicates present climate projections, and the red to yellow lines shows future climate projections in MRI-
AGCM 3.2S (the solid line is MME SST, the long broken line is cluster 1, the short broken line is cluster 2, and the dot broken
line is cluster 3). The vertical bars represent standard deviations. In (d)–(f), the black marks indicate raingauge, and the red
and green marks indicates present climate projections of MRI-AGCM 3.2S and 3.2H, respectively
Figure 2. Basin average rainfall of the four major recent ood events from raingauge observation (upper row) and WRF sim-
ulation (bottom row)
2012). All the rainfall under the present and future climate
show similar variation with each other, and are consistent
with APHRODITE rainfall (black line). In Figure 3b, on the
other hand, the downscaled rainfall is overestimated com-
pared to observed rainfall. Figure 3c shows rainfall with
biases corrected based on raingauge observation, as
DYNAMIC DOWNSCALING OF PAMPANGA BASIN
—109—
explained below. The black line in Figure 3c is a basin aver-
age rainfall based on the raingauge, which is slightly smaller
than the one in Figure 3a by APHRODITE, but shows quite
similar variation. It was thus conrmed that the raingauge
observation is equivalent with APHRODITE.
The dynamic downscaling dramatically improved the
appearance of heavy rainfall from GCM, as illustrated in
Figures 3d and 3e. Figure 3d shows that the GCM-based
rainfall was quite a poor reproduction of frequency of
appearance from raingauge data. However, Mizuta et al.
(2012) have already pointed out the possibility of GCM’s
poor performance particularly in its application to tropical
western Pacic islands (see Figure 16e in their study). The
poor performance by MRI-AGCM in the case of the
Philippine islands could be interpreted in this respect.
The downscaling overestimated the number of weak rain-
fall events with less than about 100 mm day–1 (Figure 3e),
compared with raingauge observation. This overestimation
of weak rainfall was a reason for overestimated monthly
rainfall in Figure 3b. To correct this disagreement, we
employed bias correction based on popular distribution
mapping (De Troch et al., 2013). As shown in Rasmy et al.
(2014), distribution mapping alone is appropriate for bias
correction of downscaled rainfall, although a 3-step bias cor-
rection works better for GCM rainfall (Jaranilla-Sanchez et
al., 2013; Nyunt et al., 2014). Because of the limited number
of raingauge stations and incomplete data continuity, we
mapped the downscaled basin-average rainfall onto basin
average raingauge rainfall. The cumulative rainfall probabil-
ity was used without any tting on a Gamma function, since
this non-parametric method performs the best in reproduc-
ing extreme rainfall. The mapping was performed for each
month, but the top 0.5% rainfall events were mapped for
each year since annual maximum extreme rainfall can appear
in any month of the year (Inomata et al., 2009). The obtained
correction rate was multiplied by the downscaled rainfall,
and spatial distributions were recovered. The correction
rates used for the present climate were also applied to future
climate rainfall. After bias correction, seasonal variations
(Figure 3c) agreed well with observations, and the frequency
appearance also showed consistency with raingauge obser-
vation (Figure 3f).
Figure 3c compares the seasonal variation of rainfalls
from observation and downscaling of present and future cli-
mate projections. Rainfall in the wet season (May to
September), especially cluster 2 SST (orange short broken
line), slightly increased due to global warming, but the
increase was not signicant. Cluster 2 SST is an El Niño-
type pattern (Kitoh and Endo, 2016), in which SST cools
slightly around the Philippines. Although it was not clear
why cluster 2 SST projected the largest rainfall, we also
found this tendency in other analyses. As Figure 3 shows,
only a slight change in monthly rainfall is projected in the
Pampanga River Basin due to global warming.
Figure 4 indicates the results from frequency analysis of
extreme rainfall. We chose the annual maximum 48 h rain-
fall for this analysis, since our ood inundation simulation
area had the highest correlation with maximum 48 h rainfall.
Figure 4a shows that the frequency distributions of rain-
gauge rainfall, ERA-interim, and the present climate of
MRI-AGCM 3.2S were consistent with each other. In Figure
4b, the present climate of MRI-AGCM 3.2H estimated a
slightly larger rainfall for longer return periods without a
signicant degree of overestimation. This indicates that the
bias corrected downscaled rainfall events represent the pres-
ent climate rainfall features successfully.
Figure 4. The frequency analysis of annual maximum 48 h rainfall from downscaling and bias corrected results of (a) MRI-
AGCM 3.2S and (b) MRI-AGCM 3.2H. The lines are Gumbel distribution tting. The points and lines in black are raingauge
observations, the green ones are downscaled ERA-interim, the blue ones are present climate projections, the red ones are
future climate projections with MME SST, the orange long broken line is future climate projection with cluster 1 SST, the
orange short broken line is cluster 2 SST, and the orange dot-broken line is future climate projection with cluster 3 SST
T. USHIYAMA ET AL.
—110—
In Figure 4, the four lines of red to orange colors show the
frequency distributions under the future climate. The fre-
quency distributions with MME and cluster 2 SST in Figure
4a and MME SST in Figure 4b were signicantly larger than
those of the present climate. In the case of the 50-year return
period, rainfall is estimated to be 320 mm in the present cli-
mate, while it is projected to increase by 45% to 470 mm
(MRI-AGCM 3.2S, cluster 2) and by 65% to 530 mm (MRI-
AGCM 3.2H) in the future climate. The results suggest a
drastic increase in extreme rainfall in the Pampanga River
Basin due to global warming. However, the two lines in
MRI-AGCM 3.2S (clusters 1 and 3) in Figure 4a show little
change due to global warming, suggesting uncertainty in the
range of increase.
To conrm the credibility of the extreme rainfall changes
under the future climate, we further examined those amounts
by original MRI-AGCM 3.2S and 3.2H without downscal-
ing and without bias correction in the same way as in Figure
4 (gures are not shown). The maximum 48 h rainfalls
obtained in the 50-year return period were 329 mm and
317 mm under the present climate for MRI-AGCM 3.2S and
3.2H, respectively, and they increased to 551 mm and
461 mm, respectively, under the RCP8.5 scenario in MME
SST. The rates of increase in extreme rainfall due to global
warming by original MRI-AGCM 3.2S and 3.2H were 67%
and 45%, respectively. Thus, the rates of increase in extreme
rainfall, 45% and 65%, obtained from the downscaled rain-
fall are reasonable based on MRI-AGCM 3.2S and 3.2H.
Our analysis suggests a potential increase in extreme rain-
fall due to global warming. This is consistent with past stud-
ies (IPCC, 2013; Jaranilla-Sanchez et al., 2013; Kitoh and
Endo, 2016). The advantage of this study is the use of a down-
scaling model with a higher resolution. Jaranilla-Sanchez
et al. (2013) showed that, based on bias-corrected rainfall
from six GCMs, the 10th percentile peak daily discharge
may increase in the Pampanga River Basin from the years of
1981–2000 to 2046–2065 in the range of zero to six times. It
is not appropriate to directly compare their results with ours.
However, our analysis using a downscaling approach has
contributed to reduction of the uncertainty concerning a
potential increase in extreme rainfall due to global warming.
To discuss the mechanism of rainfall increase due to
global warming, monsoon strength was compared between
the present and future climate in Figure 5. The monsoon
trough along 20°N is dominant in the present climate, driv-
ing the south-westerly monsoon in the summer time (Figure
5a). In contrast, the monsoon trough may weaken in the
future climate, resulting in a weaker monsoon ow (Figure
5b) and a resultant decrease in west coast rainfall (gure not
shown). The weakening of monsoon circulation is also men-
tioned in the IPCC AR5 report (IPCC, 2013). All this implies
that, at least, the average monsoon ow may not strengthen
the projected extreme rainfall.
In this river basin, the forcing of extreme rainfall is due to
either typhoons or monsoon. Figure 6 compares the causes
of annual maximum rainfall in each year in each scenario,
which was determined subjectively based on temporal vari-
ation of rainfall distributions. Extreme rainfalls can be
attributed to typhoons in 27 out of 50 years (54%) between
1979 and 2003 in the present climate, while in 30 out of 125
years (24%) between 2075 and 2099 in the future climate.
This implies that the number of typhoon-induced rainfall
events may signicantly decrease due to global warming.
This implication is consistent with many previous studies
(e.g., Murakami et al., 2012) reporting a potential decrease in
the number of typhoons in the Western Pacic due to global
warming, although the number of very strong typhoons is
projected to increase. Nevertheless, very intense extreme
rainfall events tend to be attributed to typhoons in most
cases, as Figure 6 indicates, which suggests that typhoon
rainfall would still remain important in forcing very extreme
rainfall in the future climate.
Figure 5. Sea level pressure (contours; unit: hPa) and surface wind vectors in August of the present (a) and future (b) climate
(MME SST) in MRI-AGCM 3.2S
DYNAMIC DOWNSCALING OF PAMPANGA BASIN
—111—
SUMMARY AND CONCLUSIONS
We investigated the impact of global warming on changes
in extreme rainfall in a ood prone river basin of the
Philippines under the present (1979–2003) and future
(2075–2099) climate, using dynamic downscaling of MRI-
AGCM 3.2S and 3.2H and the RCP 8.5 scenario. Since
downscaled rainfall overestimates the number of weak rain-
fall events, bias correction was applied based on distribution
mapping with data from raingauge observations. The cor-
rected present climate rainfall showed reasonable agreement
with raingauge observation with regard to monthly rainfall,
frequency appearance of daily rainfall, and frequency analy-
sis. The monthly rainfall in the future climate did not show
any signicant changes from that in the present climate.
However, the frequency analysis found that extreme rainfall
may increase in the future climate, depending on SST condi-
tions. With four SST conditions in MRI-AGCM 3.2S and
one SST condition in MRI-AGCM 3.2H, future climate pro-
jections suggested a possible increase with large uncertainty,
having a wide range from 0 to 65%. The increase in extreme
rainfall projected in this study is consistent with past studies.
In addition, the study contributed to specifying an increase
in range. This rainfall data will be used in future hydrologi-
cal simulations to examine changes in ood inundation
extent due to global warming.
Extreme rainfall events over the Philippines are caused by
either typhoons or monsoon ows. The strength of the East
Asian monsoon in summer may decrease in the future. In
contrast, 54% of annual extreme rainfall events were attri-
butable to typhoons in the present climate, compared to only
24% in the future climate. This means that the monsoon may
become weaker, and the number of typhoons may decrease,
in the future climate. Further study is necessary to elucidate
the mechanism of a potential increase in extreme rainfall.
ACKNOWLEDGMENTS
This work was conducted under the Program for Risk
Information on Climate Change supported by the Ministry
of Education, Culture, Sports, Science, and Technology-
Japan (MEXT). The Grid Analysis and Display System
(GrADS) and GFD Dennou Club Library were used to draw
gures.
REFERENCES
Chotamonsak C, Salathe Jr EP, Kreasuwan J, Chantara S,
Siriwitayakorn K. 2011. Projected climate change over
Southeast Asia simulated using a WRF regional climate
model. Atmospheric Science Lettters 12: 213–219. DOI:
10.1002/asl.313.
Cruz RT, Sasaki H, Narisma GT. 2016. Assessing the sensitivity of
the non-hydrostatic regional climate model to boundary con-
ditions and convective schemes over the Philippines. Journal
Figure 6. Annual maximum 48 h rainfall. The yellow bars indicate typhoon attributions and the green bars indicate monsoon
attributions. The two panels in the blue rectangle are from the present climate, and the ve panels in the red rectangle are from
the future climate
T. USHIYAMA ET AL.
—112—
of the Meteorological Society of Japan 91A: 165–179. DOI:
10.2151/jmsj.2015-059.
De Troch R, Hamdi R, HV de Vyver, Geleyn JF, Termonia P. 2013.
Multiscale performance of the ALARO-0 model for simulat-
ing extreme summer precipitation climatology in Belgium.
Journal of Climate 26: 8895–8915. DOI: 10.1175/JCLI-
D-12-00844.1.
Grell GA, Devenyi D. 2002. A generalized approach to parameter-
izing convection combining ensemble and data assimilation
techniques. Geophysical Research Letters 29: 38-1–38-4.
DOI: 10.1029/2002GL015311.
Inomata H, Takeuchi K, Fukami K. 2009. A study on a statistical
bias correction method for GCM rainfall data. Annual Journal
of Hydraulic Engineering, Japan Society of Civil Engineers
53: 223–228 (in Japanese with English abstract).
Intergovernmental Panel on Climate Change (IPCC). 2013. The
physical science basis in the fth assessment report of the
intergovernmental panel on climate change. IPCC. Cambridge
University Press, Cambridge, UK; 1535.
Jaranilla-Sanchez RA, Koike T, Nyunt CT, Rasmy M, Hasegawa I,
Matsumura A, Ogawa D. 2013. Hydrological impacts of a
changing climate on oods and droughts in Philippine river
basins. Journal of Japan Society of Civil Engineers, Ser. B1
(Hydraulic Engneering) 69: I_13–I_18.
Jose AM, Cruz NA. 1999. Climate change impacts and responses in
the Philippines: water resources. Climate Research 12: 77–84.
DOI: 10.3354/cr012077.
Kain JS, Fritsch JM. 1993. Convective parameterization for meso-
scale models: The Kain-Fritsch scheme. In The representation
of cumulus convection in numerical models. Emanuel KA,
Raymond DJ (eds). Meteorological Monographs 24:
American Meteorological Society, Boston, USA; 165–170.
DOI: 10.1007/978-1-935704-13-3_16.
Kanada S, Nakano M, Hayashi S, Kato T, Nakamura M, Kurihara
K, Kitoh A. 2008. Reproducibility of maximum daily
precipitation amount over Japan by a high-resolution non-
hydrostatic model. SOLA 4: 105–108. DOI: 10.2151/sola.
2008‒027.
Kitoh A, Endo H. 2016. Changes in precipitation extremes pro-
jected by a 20-km mesh global atmospheric model. Weather
and Climate Extremes 11: 41–52. DOI: 10.1016/j.wace.2015.
09.001.
Mizuta R, Yoshimura H, Murakami H, Matsueda M, Endo H, Ose
T, Kamiguchi K, Hosaka M, Sugi M, Yukimoto S, Kusunoki
S, Kitoh A. 2012. Climate simulations using MRI-AGCM3.2
with 20-km grid. Journal of the Meteorological Society of
Japan 90A: 233–258. DOI: 10.2151/jmsj.2012-A12.
Murakami H, Wang Y, Yoshimura H, Mizuta R, Sugi M, Shindo E,
Adachi Y, Yukimoto S, Hosaka M, Kusunoki S, Ose T, Kitoh
A. 2012. Future changes in tropical cyclone activity projected
by the new high-resolution MRI-AGCM. Journal of Climate
25: 3237–3260. DOI: 10.1175/JCLI-D-11-00415.1.
Narita M. 2008. Improvement of Kain-Fritsch scheme and parame-
ter tuning. Report of Numerical Prediction division, Japan
Meteorological Agency, Extra Volume 54: 103–111 (in
Japanese).
Nyunt CT, Koike T, Yamamoto A, Nemoto T, Kitsuregawa M.
2014. Application of statistical bias correction method to the
Yoshino river basin. Journal of Japan Society of Civil
Engineers, Ser. B1 (Hydraulic Engneering) 70: I_193–I_198.
Rasmy M, Shrestha M, Koike T, Hara M, Fujita M, Kimura F.
2014. A combined dynamical/statistical downscaling
approach for assessing future of water resources in the Tone
river basin. Journal of Japan Society of Civil Engineers, Ser.
B1 (Hydraulic Engneering) 70: I_187–I_192.
Shrestha BB, Okazumi T, Miyamoto M, Sawano H. 2014.
Development of ood risk assessment method for data-poor
river basins: a case study in the Pampanga river basin,
Philippines. Proceedings of 6th International Conference on
Flood Management, September 16–18, 2014, Sao Paulo,
Brazil, http://www.abrh.org.br/icfm6/proceedings/papers/
PAP014383.pdf. Last access August 30, 2016.
Skamarock WC, Klemp JB, Dudhia J, Gill DO, Barker DM, Wang
W, Powers JG. 2005. A description of the Advanced Research
WRF version 2. NCAR Technical Note TN-468+STR,
Mesoscale Meteorology Division, National Center for
Atmospheric Research, Colorado, USA; 88.
Wang D, Menz C, Simon T, Simmer C, Ohlwein C. 2013. Regional
dynamical downscaling with CCLM over East Asia.
Meteorology and Atmospheric Physics 121: 39–53. DOI:
10.1007/s00703-013-0250-z.
Yatagai A, Kamiguchi K, Arakawa O, Hamada A, Yasutomi N,
Kitoh A. 2012. APHRODITE: Constructing a long-term daily
gridded precipitation dataset for Asia based on a dense net-
work of rain gauges. Bulletin of the American Meteorological
Society 93: 1401–1415. DOI: 10.1175/BAMS-D-11-00122.1.
... MRI-AGCM3.2S, a high-resolution 20-km grid model, was produced by the Meteorological Research Institute of Japan (Mizuta et al. 2012;Kitoh and Endo 2016), and the model was designed to generate 20 km fine resolution outputs, much finer resolution outputs compared with those produced by other Coupled Model Intercomparison Project Phase 5 (CMIP5) general circulation models (GCMs) (Mizuta et al. 2012 performed with grid-based daily APHRODITE precipitation data (Asian Precipitation-Highly Resolved Observational Data Integration toward Evaluation of Water Resources-APHRODITE), and the same bias correction factors were also applied to the future climate case . GCM data were downscaled only for PRB using a dynamic downscaling method since the flood travel time is short on hourly basis in this basin (Ushiyama et al. 2016;Iwami et al. 2017). The bias correction of MRI-AGCM3.2S ...
... The bias correction of MRI-AGCM3.2S for the study river basins and also the downscaling of GCM data for PRB can be found in detail in Ushiyama et al. (2016) and Iwami et al. (2017). ...
Article
Full-text available
This study focused on flood damage assessment for future floods under the impact of climate change. Four river basins of Southeast Asia were selected for the study. They included the Pampanga River Basin (PRB) in the Philippines, the Solo River Basin (SRB) in Indonesia, the Lower Mekong River Basin (LMRB) in Cambodia and Vietnam, and the Chao Phraya River Basin (CPRB) in Thailand. Flood damage to rice crops was assessed by flood damage functions considering flood depth and duration and the growth stage of rice plants. Flood characteristics such as flood depth, duration, and distribution were computed using the rainfall–runoff–inundation model to assess flood hazards under the present and future climatic conditions produced by MRI-AGCM3.2S. The damage assessment methodology for rice crops employed in this study was verified using data on past flood events. Then, flood damage assessment was conducted for both the present climate (1979–2003) and future climate (2075–2099) conditions, using MRI-AGCM3.2S precipitation datasets. Flood damage was assessed for worst cases chosen from each climate period and for floods of 50- and 100-year return periods with different rainfall patterns chosen from each climate scenario. The results of flood hazard and damage assessment show that the flood inundation area for a 100-year flood may increase in the future by 20% in PRB; by 66% in SRB; by 27% in LMRB; and by 27% in CPRB. The flood damage area of paddy fields for a 100-year flood may also increase in the future by 16% in PRB; by 55% in SRB; by 23% in LMRB; and by 13% in CPRB.
... The RCP8.5 Scenario of MRI-AGCM 3.2H was dynamically downscaled to 5 km by the WRF model with no convection scheme. Since the probability result of past climate after downscaling showed gaps with observed data, CHIRPS was used to perform bias correction based on monthly factors of cumulative probabilities for daily rainfall, excluding the top 0.5%, as was applied to the Pampanga River basin in the Philippines [30]. Regarding the downscaled GCM outputs, the worst extreme events in the past climate from 1979 to 2003 and the future climate from 2075 to 2099 were identified in terms of 24 h maximum rainfall and used for the impact assessment of the hazard estimation. ...
Article
Full-text available
Enhancing flood resilience, including the development of social capacity and early warning systems, in addition to structural measures, is one of the key solutions to mitigating flood damage, which will be more intensified in the future due to climate change. This study was conducted to develop a comprehensive methodology for enhancing flood resilience by improving society-wide disaster literacy under the governance formed through the active participation of all levels of stakeholders in Davao City, Philippines. Specifically, the development of the Online Synthesis System for Sustainability and Resilience, which integrates different disciplines, and the fostering of Facilitators, whose role is to interlink the science community and society, were implemented in a co-designing manner by the collective governance body. The development of basin- and barangay-scale hydrological models realized real-time flood forecasting and climate change impact assessment to identify intensified flood risk under the future climate. Co-designed e-learning workshops were held to foster about thirty Facilitators and help them produce twenty-one risk communication plans and workshop designs for fourteen barangays considering geographic, demographic, economic, and social features that they can utilize for public dissemination related to climate change adaptation to the target audiences in society. This paper presents a practical method to enhance flood resilience, demonstrating that the synthesis of science-based knowledge and human resource development can fill the gaps between the science community and society.
... The next report to be released will include simulations from CORDEX-SEA and discuss future changes in climate extremes. Furthermore, other studies have conducted downscaling of climate projections for sectoral applications using dynamical (e.g., for assessment on wind energy (Silang et al. 2014) and flood risk (Ushiyama et al. 2016)) and statistical approaches (e.g., for climate change adaptation planning (Basconcillo et al. 2016)). ...
Chapter
This chapter reviews the progress in regional climate downscaling simulations (RCDS) over Southeast Asia, both at country and regional levels. The need to advance RCDS stems from the fact that robust climate change projections are needed in decision-making processes involved in adopting climate-resilient pathways, wherein adaptation, mitigation, and sustainable development aspects must also be considered. At the country level, notable variation in the progress in RCDS among countries in Southeast Asia exists wherein some countries have not conducted any RCDS, while others show some levels of activities in both dynamical and statistical downscaling. At the regional level, RCDS activities are also limited. However, the establishment of the Coordinated Regional Climate Downscaling Experiment (CORDEX) Southeast Asia (CORDEX-SEA) and the completion of high-resolution multi-model simulations are game changers. Furthermore, through the establishment of the Southeast Asia Regional Climate Change Information System (SARCCIS), a data portal where CORDEX-SEA simulation outputs are archived and linked to the Earth System Grid Federation (ESGF) for worldwide accessibility, the availability of robust climate change projections is expected to spur the research and development activities in the vulnerability, impact, and adaptation aspects in the region. Lastly, the remaining challenges, such as on bias and uncertainties, are discussed, which indicate the way forward concerning RCDS in Southeast Asia.
... In general, temporal and spatial downscaling is necessary for a small basin where the flood duration is short on an hourly basis and floods are influenced by microtopography . We performed a dynamic downscaling for the Pampanga river basin using Weather Research and Forecasting (WRF) model with 5 km grid on an hourly basis to prepare input precipitation data for flood runoff and inundation calculation (Ushiyama et al., 2016). On the other hand, we did not conduct downscaling for the Solo river basin despite that it is a small basin, because the flood duration was about 4 days and we were able to reproduce a hydrograph on a daily basis. ...
Article
As many water related disasters occur frequently around the world, proper assessment of future climate change impact on floods and droughts is essential. In this study, we focused on basin-scale climate change impact assessment as necessary information for studying adaptation measures on the basis of integrated water resources management. We used Meteorological Research Institute-Atmospheric General Circulation Model (MRI-AGCM) 3.2S (20km grid super high resolution model) and a series of simulation methods for climate change analysis. We conducted a comparative study on changes in precipitation, flood discharge and inundation in the future during the wet and dry seasons for five target river basins in the Asian monsoon area. We found that regional precipitation outputs from the high resolution model in this study were in good agreement in the point of tendency of their changes in wet and dry monsoon seasons with the regional precipitation analysis in the Fifth Assessment Report (AR5) of Working Group 1 (WG1) of the International Panel on Climate Change (IPCC, 2013). This study illustrated that the proposed methodology can make more detailed descriptions of climate change possible. The study also found the importance of basin-scale runoff and inundation analysis with downscaling especially for basins where floods occur for a short period, suggesting potential differences between flood change and precipitation change from General Circulation Model (GCM) outputs such as maximum 5-day precipitation index in a basin. As a result, this paper confirms the importance of basin-scale discharge and inundation analysis for climate change considering basin characteristics from the viewpoint of river management.
Article
Full-text available
High-resolution modeling is necessary to project weather and climate extremes and their future changes under global warming. A global high-resolution atmospheric general circulation model with grid size about 20km is able to reproduce climate fields as well as regional-scale phenomena such as monsoonal rainfall, tropical and extratropical cyclones, and heavy precipitation. This 20-km mesh model is applied to project future changes in weather and climate extremes at the end of the 21st century with four different spatial patterns in sea surface temperature (SST) changes: one with the mean SST changes by the 28 models of the Coupled Model Intercomparison Project Phase 5 (CMIP5) under the Representative Concentration Pathways (RCP)-8.5 scenario, and the other three obtained from a cluster analysis, in which tropical SST anomalies derived from the 28 CMIP5 models were grouped. Here we focus on future changes in regional precipitation and its extremes. Various precipitation indices averaged over the twenty-two regional land domains are calculated. Heavy precipitation indices (maximum 5-day precipitation total and maximum 1-day precipitation total) increase in all regional domains, even where mean precipitation decrease (Southern Africa, South Europe/Mediterranean, Central America). South Asia is the domain of the largest extreme precipitation increase. In some domains, different SST patterns result in large precipitation changes, possibly related to changes in large-scale circulations in the tropical Pacific.
Conference Paper
Full-text available
In recent years, floods frequently occur in the world, which cause serious damage of properties and loss of lives. An assessment of flood risk in the river basin is thus very important for development activities and river basin management. In this study, the flood damage assessment method was developed for the assessment of flood risk in data-poor river basins by using physical based numerical model, satellite-based information and socio-economic factors. The Pampanga river basin of the Philippines was selected for the case study. For the flood damage assessment, the hazard characteristics such as flood depth and duration were computed by using Rainfall Runoff Inundation (RRI) model. Agriculture and household which are major exposures in the flood prone areas were taken into account for the flood damage assessment. The potential damages of agriculture were estimated by using the damage function of agriculture and flood characteristics. The damage functions of agriculture for its each growing stages were defined as the function of flood depth and duration. Field investigation and questionnaire surveys were conducted at the selected Barangays (Villages) in the Pampanga river basin to develop methodology for household damage estimation. The appropriate damage functions for household building and assets damages were developed based on the flood and household characteristics. The estimated damages of agriculture and household were consistent with the reported values. The developed method can be applied in other river basins where damages data are poorly available.
Article
Full-text available
A new version of the atmospheric general circulation model of the Meteorological Research Institute (MRI), with a horizontal grid size of about 20 km, has been developed. The previous version of the 20-km model, MRIAGCM3.1, which was developed from an operational numerical weather-prediction model, provided information on possible climate change induced by global warming, including future changes in tropical cyclones, the East Asian monsoon, extreme events, and blockings. For the new version, MRI-AGCM3.2, we have introduced various new parameterization schemes that improve the model climate. Using the new model, we performed a present-day climate experiment using observed sea surface temperature. The model shows improvements in simulating heavy monthly-mean precipitation around the tropical Western Pacific, the global distribution of tropical cyclones, the seasonal march of East Asian summer monsoon, and blockings in the Pacific. Improvements in the model climatologies were confirmed numerically using skill scores (e.g., Taylor’s skill score).
Article
Full-text available
New versions of the high-resolution 20- and 60-km-mesh Meteorological Research Institute (MRI) atmospheric general circulation models (MRI-AGCM version 3.2) have been developed and used to investigate potential future changes in tropical cyclone (TC) activity. Compared with the previous version (version 3.1), version 3.2 yields a more realistic simulation of the present-day (1979-2003) global distribution of TCs. Moreover, the 20-km-mesh model version 3.2 is able to simulate extremely intense TCs (categories 4 and 5), which is the first time a global climate model has been able to simulate such extremely intense TCs through a multidecadal simulation. Future (2075-99) projections under the Intergovernmental Panel on Climate Change (IPCC) A1B scenario are conducted using versions 3.1 and 3.2, showing consistent decreases in the number of TCs globally and in both hemispheres as climate warms. Although projected future changes in basin-scale TC numbers show some differences between the two versions, the projected frequency of TC occurrence shows a consistent decrease in the western part of the western North Pacific (WNP) and in the South Pacific Ocean (SPO), while it shows a marked increase in the central Pacific. Both versions project a future increase in the frequency of intense TCs globally; however, the degree of increase is smaller in version 3.2 than in version 3.1. This difference arises partly because version 3.2 projects a pronounced decrease in mean TC intensity in the SPO. The 20-km-mesh model version 3.2 projects a northward shift in the most intense TCs (category 5) in the WNP, indicating an increasing potential for future catastrophic damage due to TCs in this region.
Article
Regional climate models have been useful in climate studies and in downscaling climate projections from global climate models, especially for areas characterized by complex topography and coastline features, such as the Philippines. However, several factors may affect model skill, such as uncertainties related to the boundary conditions and model configuration. This study evaluates the performance of the non-hydrostatic regional climate model (NHRCM) over the Philippines. Present-day climate simulations at 50 km resolution are conducted using two sets of boundary conditions (ECMWF ERA-Interim and the NCEP/NCAR Reanalysis Project NNRP1), as well as two convective parameterization schemes in the model (Grell and Kain-Fritsch). Results show that the seasonal changes in the spatial distribution of temperature, rainfall, and winds over the Philippines are simulated reasonably well. NHRCM has an overall cold and dry bias over land, the degree of which depends on the boundary condition and the convective scheme used. After adjusting the simulated temperature because of the difference in topography, the temperature differs from that observed by -0.90°C to -0.42°C on average. The rainfall bias in NHRCM ranges from -62.13 % to -25.20 %. Regardless of the boundary condition, the Grell scheme results in the lowest temperature bias with high skill scores, while the Kain-Fritsch scheme gives the lowest rainfall bias with high correlation and skill scores. The boundary conditions also influence model skill, such that the model bias is lower for temperature when ERA-Interim is used, but lower for rainfall with NNRP1. NHRCM represents the seasonal cycles of temperature and rainfall for all regions, but tends to generate more occurrences of cold and dry months. Improvements in the model are still possible, but these results indicate the potential of the model to be used for providing essential information for describing historical and future changes in the Philippine climate.
Article
Dynamical downscaling is a promising tool to assess the future fate of water in a catchment. To overcome strong biases in Coupled Ocean-Atmosphere General Circulation Models (CGCMs), the Pseudo Global Warming Downscaling (PGW-DS), which combines climatology differences (future-past) of CGCM ensembles and Reanalysis Dataset (RD) was proposed as a reliable and efficient method, but downscaling of RD at a basin scale was not assessed. Hence, this study evaluated the results from a long-term, high-resolution Downscaled Precipitation (DP) derived from a RD over the Tone river basin. DP showed strong bias over mountainous regions owing to resolution enhancement. DP and CGCM precipitations were merged with Statistical Bias Correction methods (SBC) to obtain a comprehensive outlook on biases and the most appropriate SBC. Characteristics of biases were quite different in DP and CGCM, which performed very poor for extreme rainfall intensities and thus required an explicit treatment. Seasonal inconsistency in extreme events of CGCMs affected the corrected rainfall and discharges significantly. Conversely, a simple cumulative gamma method was successful for DP to represent climatology, extreme statistics of rainfall and discharges owing to the use of RD, which generated lesser bias compared to CGCMs. The proposed method will be applied to PGW-DS to obtain reliable information of future changes of water in the basin.
Article
Considering still high uncertainty within results of climate projections by General Circulation Models (GCMs) with various emission scenarios, a multi-model approach after bias corrections and downscaling is very important for climate change impact assessment on water resources at a river basin scale. In addition to the methodologies for selecting GCMs which express regional climate characteristics reasonably, this paper aims to reduce the GCM bias of extreme rainfall where most of GCMs underestimate and improve the poor seasonal variation of GCM bias too. In the Yoshino River, where heavy rainfall events are mainly characterized by the Baiu in June and July and typhoon in August and September, we recognized big bias in extremely heavy rainfalls from June to September. Investigating the tendency of GCMs, seasonal bias corrections were introduced additionally. The results of holistic bias correction and downscaling provide reliable, quantitative and qualitative information to the policy makers for planning for integrated water management in the Yoshino River.