IGAD Climate Predictions and Applications Centre Technical Report 20/01/06/2009 ASSESSMENT OF THE SKILL/ACCURACY OF THE REGIONAL SPECTRAL MODEL (RSM) OVER THE GHA REGION Verification Analysis for Seasonal Temperature and Rainfall Prediction in the GHA Region Using RSM Franklin J. Opijah, Joseph M. Mutemi
57
Embed
IGAD Climate Predictions and Applications Centre - ICPACrcc.icpac.net/wp-content/uploads/...and-Rainfall-in-the-GHA-Region.pdf · IGAD Climate Predictions and Applications Centre
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
IGAD Climate Predictions and Applications Centre Technical Report
20/01/06/2009
ASSESSMENT OF THE SKILL/ACCURACY OF THE REGIONAL SPECTRAL MODEL (RSM) OVER THE GHA
REGION
Verification Analysis for Seasonal Temperature and Rainfall Prediction in the GHA Region Using RSM
a. The Main Precipitating Weather Systems over the GHA Region ..................... 1 b. Forcing of Organized Convection ...................................................................... 1 c. The Regional Spectral Model .......................... Error! Bookmark not defined.
2. Verification Methodology Applied ........................................................................ 2 a. Simple Correlation and Error Analysis .............................................................. 5 b. Contingency Tables and Skill Scores ................................................................. 5
3. RSM Performance over Kenya .............................................................................. 7 a. RSM Temperature Verification over Kenya for OND Season .......................... 7 b. RSM Rainfall Verification over Kenya for OND Season .................................. 9 c. RSM Temperature Verification over Kenya for MAM Season ....................... 11 d. RSM Rainfall Verification over Kenya for MAM Season .............................. 13 e. RSM Temperature Verification over Kenya for JJA Season ........................... 15 f. RSM Rainfall Verification over Kenya for the JJA Season ............................. 17
4. RSM Performance over Tanzania ........................................................................ 19 a. RSM Rainfall Verification over Tanzania for OND Season ............................ 19 b. RSM Rainfall Verification over Tanzania for MAM Season .......................... 22 c. RSM Rainfall Verification over Tanzania for JJA Season .............................. 24
5. RSM Performance over Burundi, Eritrea and Djibouti ........................................ 27 a. RSM Rainfall Verification over Burundi, Eritrea and Djibouti for OND Season 27 b. RSM Rainfall Verification over Burundi, Eritrea and Djibouti for MAM Season ...................................................................................................................... 29 c. RSM Rainfall Verification over Burundi, Eritrea and Djibouti for JJA Season 31
6. RSM Performance over Ethiopia ......................................................................... 33 a. RSM Rainfall Verification over Ethiopia for OND Season ............................. 33 b. RSM Rainfall Verification over Ethiopia for MAM Season ........................... 35 c. RSM Rainfall Verification over Ethiopia for JJA Season ............................... 37
7. RSM Performance over Sudan ............................................................................. 39 a. RSM Rainfall Verification over Sudan for OND Season ................................ 39 b. RSM Rainfall Verification over Sudan for MAM Season ............................... 41 c. RSM Rainfall Verification over Sudan for JJA Season ................................... 44 d. RSM Rainfall Verification over Sudan for JJA Season ................................... 47
8. Dynamical Predictability over the Greater Horn of Africa .................................. 50 a. Degree of Predictability over Greater Horn of Africa ..................................... 50 b. The Challenge of Limitation to Predictability ................................................. 51
Figure 1: Forecasted seasonal rainfall (left) and temperature (right) over the GHA region using RSM and ECHAM for JJA, 2008 ......................................................................... 3
Figure 2: Difference fields (anomalies) for OND, 2008. .................................................... 3 Figure 3: Forecast Difference fields (anomalies) for MAM, 2009. .................................... 4 Figure 4: Inter-annual temperature variability over Wajir, Kenya during the OND season.
........................................................................................................................................ 8 Figure 5: Inter-annual temperature variability over Mombasa, Kenya during the OND
season. ........................................................................................................................... 9 Figure 6: Inter-annual rainfall variability over Nyeri, Kenya during the OND season. .. 10 Figure 7: Inter-annual rainfall variability over Malindi, Kenya during the OND season.
...................................................................................................................................... 11 Figure 8: Inter-annual temperature variability over Kitale, Kenya during the MAM
season. ......................................................................................................................... 12 Figure 9: Inter-annual temperature variability over Kisumu, Kenya during the MAM
season. ......................................................................................................................... 13 Figure 10: Inter-annual rainfall variability over Kisii, Kenya during the MAM season. 14 Figure 11: Inter-annual rainfall variability over Nakuru, Kenya during the MAM rainfall
season. ......................................................................................................................... 15 Figure 12: Inter-annual temperature variability over Kitale, Kenya during the JJA season.
...................................................................................................................................... 16 Figure 13: Inter-annual temperature variability over Kisumu, Kenya during the JJA
season. ......................................................................................................................... 17 Figure 14: Inter-annual temperature variability over Dagoretti, Kenya during the JJA
season. ......................................................................................................................... 17 Figure 15: Inter-annual rainfall variability over Nakuru, Kenya during the MAM rainfall
season. ......................................................................................................................... 18 Figure 16: Inter-annual rainfall variability over Nakuru, Kenya during the MAM rainfall
season. ......................................................................................................................... 19 Figure 17: Inter-annual rainfall variability over Dodoma, Tanzania during the OND
rainfall season. ............................................................................................................ 20 Figure 18: Inter-annual rainfall variability over Iringa, Tanzania during the OND rainfall
season. ......................................................................................................................... 21 Figure 19: Inter-annual rainfall variability over Dar es Salaam International Airport,
Tanzania during the OND rainfall season. .................................................................. 21 Figure 20: Inter-annual rainfall variability over Mtwara, Tanzania during the OND
rainfall season. ............................................................................................................ 22 Figure 21: Inter-annual rainfall variability over Dodoma, Tanzania during the MAM
rainfall season. ............................................................................................................ 23 Figure 22: Inter-annual rainfall variability over Mwanza, Tanzania during the MAM
rainfall season. ............................................................................................................ 24 Figure 23: Inter-annual rainfall variability over Mtwara, Tanzania during the JJA rainfall
season. ......................................................................................................................... 25 Figure 24: Inter-annual rainfall variability over Mbeya, Tanzania during the JJA rainfall
season. ......................................................................................................................... 26 Figure 25: Inter-annual rainfall variability over Kigoma, Tanzania during the JJA rainfall
season. ......................................................................................................................... 26 Figure 26: Inter-annual rainfall variability over Bukoba, Tanzania during the JJA rainfall
season. ......................................................................................................................... 27 Figure 27: Inter-annual rainfall variability over Bujumbura, Burundi during the OND
Figure 28: Inter-annual rainfall variability over Djibouti, Djibouti during the OND rainfall season. ............................................................................................................ 28
Figure 29: Inter-annual rainfall variability over Asmara, Eritrea during the OND rainfall season. ......................................................................................................................... 29
Figure 30: Inter-annual rainfall variability over Bujumbura, Burundi during the MAM rainfall season. ............................................................................................................ 30
Figure 31: Inter-annual rainfall variability over Djibouti, Djibouti during the MAM rainfall season. ............................................................................................................ 30
Figure 32: Inter-annual rainfall variability over Asmara, Eritrea during the MAM rainfall season. ......................................................................................................................... 31
Figure 33: Inter-annual rainfall variability over Bujumbura, Burundi during the JJA rainfall season. ............................................................................................................ 32
Figure 34: Inter-annual rainfall variability over Djibouti, Djibouti during the JJA rainfall season. ......................................................................................................................... 32
Figure 35: Inter-annual rainfall variability over Asmara, Eritrea during the JJA rainfall season. ......................................................................................................................... 33
Figure 36: Inter-annual rainfall variability over Nekemte, Ethiopia during the OND rainfall season. ............................................................................................................ 34
Figure 37: Inter-annual rainfall variability over Combolcha, Ethiopia during the OND rainfall season. ............................................................................................................ 34
Figure 38: Inter-annual rainfall variability over Debrezeit, Ethiopia during the MAM rainfall season. ............................................................................................................ 36
Figure 39: Inter-annual rainfall variability over Gonder, Ethiopia during the MAM rainfall season. ............................................................................................................ 36
Figure 40: Inter-annual rainfall variability over Baherdar, Ethiopia during the JJA rainfall season. ............................................................................................................ 38
Figure 41: Inter-annual rainfall variability over Jima, Ethiopia during the JJA rainfall season. ......................................................................................................................... 38
Figure 42: Inter-annual rainfall variability over Port Sudan, Sudan during the OND rainfall season. ............................................................................................................ 40
Figure 43: Inter-annual rainfall variability over Kassala, Sudan during the OND rainfall season. ......................................................................................................................... 41
Figure 44: Inter-annual rainfall variability over Wau, Sudan during the OND rainfall season. ......................................................................................................................... 41
Figure 45: Inter-annual rainfall variability over Kadugli, Sudan during the OND rainfall season. ......................................................................................................................... 41
Figure 46: Inter-annual rainfall variability over Juba, Sudan during the MAM rainfall season. ......................................................................................................................... 43
Figure 47: Inter-annual rainfall variability over Kadugli, Sudan during the MAM rainfall season. ......................................................................................................................... 43
Figure 48: Inter-annual rainfall variability over Nyala, Sudan during the MAM rainfall season. ......................................................................................................................... 44
Figure 49: Inter-annual rainfall variability over Sennar, Sudan during the MAM rainfall season. ......................................................................................................................... 44
Figure 50: Inter-annual rainfall variability over Wad Halfa, Sudan during the JJA rainfall season. ......................................................................................................................... 46
Figure 51: Inter-annual rainfall variability over Malakal, Sudan during the JJA rainfall season. ......................................................................................................................... 46
Figure 52: Inter-annual rainfall variability over Wau, Sudan during the JJA rainfall season. ......................................................................................................................... 47
Figure 53: Inter-annual rainfall variability over Kassala, Sudan during the JJAS rainfall season. ......................................................................................................................... 48
iv
Figure 54: Inter-annual rainfall variability over El Geneina, Sudan during the JJAS rainfall season. ............................................................................................................ 49
Figure 55: Inter-annual rainfall variability over Wad Medani, Sudan during the JJAS rainfall season. ............................................................................................................ 49
Figure 56: Inter-annual rainfall variability over Wau, Sudan during the JJAS rainfall season. ......................................................................................................................... 50
Figure 57: Inter-annual rainfall variability over Abu Naama, Sudan during the JJAS rainfall season. ............................................................................................................ 50
1
1. Introduction
As a whole, Eastern Africa, including the Greater Horn of Africa (GHA) that
encompasses the Intergovernmental authority on Development (IGAD) countries
together with East Africa region, the Western Indian Ocean, and North-eastern parts
of the Southern Africa Development Corporation (SADC), has varied weather and
climate controls.
The objective of this research report was to assess the performance of the
Regional Spectral Model (RSM) in downscaling Echam model outputs over the GHA
region, with respect to rainfall and temperature prediction.
a. The Main Precipitating Weather Systems over the GHA Region
Thunderstorms within the ITCZ, more commonly as ordinary isolated thunderstorms,
but occasionally as super cell storm-type, are the main precipitating systems over the
Greater Horn of Africa during the main rainfall seasons. Multi-cell storms occur
mainly during non-seasonal rainfall events. Although tropical cyclones mostly affect
countries in south-eastern Africa during the southern summer, with devastating
consequences, these systems also affect parts of the Greater Horn of Africa region by
modifying the pressure and wind patterns; they are associated with depressed
precipitation over certain areas, but heavy continuous rainfall over some other parts.
Squall-systems have also been reported in portions neighbouring Lake Victoria,
although these are not very well documented. The Indian Ocean Monsoons are not
associated with heavy rainfall activity over the Equatorial East Africa, but Northeast
monsoons may have a contribution to the storms and torrential rainfall events that
affect parts of SADC during December. The Congo Air Mass is a major player in the
observed rainfall over the region, especially during the Northern Hemisphere summer
season. Along the Congo air boundary, rainfall is experienced over the wind-ward
side and very little on the leeward side of the Ethiopian highlands.
b. Forcing of Organized Convection
Convection over the Greater Horn of Africa region peaks at about midday, resulting
into afternoon convective rainfall over most places. The interaction between synoptic
scale trade winds and mesoscale circulations in form of sea-breeze circulation and
topographically induced local circulations are an important player in determining the
nature, intensity and duration of the observed diurnal cycle of precipitation.
2
Organized convective cloud systems stem from:
o Scale-interaction involving mainly the north-east and south-east trade winds and
Congo air mass, with thermally-generated mesoscale circulations induced by
land-water contrasts, particularly the Lake Victoria trough, and orographically-
induced convergence over parts of the GHA.
o Weather systems associated with low-level jet streams, particularly the wind
channelling-induced Turkana jet stream and its impacts over the Marsabit/
Turkana region of Kenya, as well as the EALLJ associated with subsident and
diffluent flow and light rainfall over the Eastern Africa region. The TEJ flows
over Northern parts of the GHA and may exert some control on surface weather
over countries like Ethiopia and Sudan.
o Equator-ward displacement of extra-tropical weather systems in certain portions
of the GHA region during winter, particularly those associated with westerlies,
blocking highs and migratory systems of the mid-latitudes
o Squall-type cloud systems over the Lake Victoria region driven by sea-breeze
circulations.
2. Seasonal Rainfall and Temperature Prediction Using RSM
The RSM
Figure 1 shows the predicted rainfall and temperature distributions over the
GHA region during the June-August season of 2008, obtained using the Regional
Spectral Model, and the original output from the global model, Echam. By and large,
the model predicted little rainfall over the region. Besides, the ITCZ that is by this
time displaced northward, the rainfall distribution was influenced by the Congo Air
mass and mesoscale features. The model predicted higher amounts of rainfall in
portions neighbouring the Lake Victoria Basin, the Ethiopian highlands and the
coastal strip. The distribution of predicted temperature was governed by the
forecasted cloud cover over the region.
3
Figure 1: Forecasted seasonal rainfall (left) and temperature (right) over the GHA region using RSM and ECHAM for JJA, 2008
Figure 2: Difference fields (anomalies) of forecasted seasonal rainfall (left) and temperature (right) over the GHA region using ECHAM and RSM for OND, 2008. The red end of the spectrum shows values above the climatological average; the blue end shows values below the average.
Figure 2 shows the predicted amounts of rainfall and temperature over the
GHA averaged for the OND, 2008, season as difference fields, where the
climatological means are subtracted from the simulated seasonal values. The diagram
shows above-normal rainfall conditions over most of the region, and below normal
temperature conditions. Evidently, simulated temperature is inversely related to cloud
cover, and hence rainfall patterns.
Figure 3 shows the forecasted total rainfall and averaged temperature
difference fields over the GHA region, for MAM, 2009. Outputs from the global
model are also shown. Both models predicted normal, tending to below normal
rainfall conditions over most of the region.
But how well does the model simulate rainfall and temperature anomalies over
the region? This question is best addressed by comparing the simulated fields with the
observed values, which must be investigated using some established criteria.
4
Figure 3: Forecast Difference fields (anomalies) of forecasted seasonal rainfall (top panel, where dark green shading shows predicted heavier rainfall; dark brown shows more arid conditions), and temperature (bottom panel, where the red end of the spectrum shows temperature values above the climatological average; the blue end shows values below the average) over the GHA region using ECHAM (Right) and RSM (Left) for MAM, 2009.
3. Verification Methodology Applied
Model verification entails the use of phrases like “correctness”, “accuracy” and
“skill”. Whereas a correct forecast is one whose outcome matches that of the forecast,
the accuracy of a forecast relates to how close the outcome of a forecast is to the
forecast itself, while the skill tells us how the forecast of one prediction method
compares to that of another method. Yet we must understand that the quality of a set
of forecasts cannot be summarised in a single number. The choice of a verification
scheme should therefore be pegged on the questions the scheme is set to address.
The overall question verification procedures attempt to address then is, is the
forecast good? The goodness of a forecast depends on the interest it generates, so a
more appropriate question would be, is the forecast good enough? The answer to this
question then relates to the application of the forecast—what the forecast is to be used
for.
Verification analysis is done using statistical techniques including correlation
and error analysis, contingency tables and graphical analysis, as described below.
5
a. Simple Correlation and Error Analyses
The simple correlation coefficient between observed and forecast fields is
computed using the following formula, where O and F are observed and forecast
fields.
( ) ( ) ⎥⎦
⎤⎢⎣
⎡−−
−−=
∑∑
∑
=
=
N
i
2i
N
1i
2i
N
1iii
FFOO
FFOOr
))((
The statistical significance of the correlation coefficient is normally assessed using the
standard t-test.
The absolute mean error (MAE) is defined as follows, where Oi and Fi are
observations and forecasts respectively, and N denotes the total number of forecasts.
∑=
−=N
1iii FO
N1MAE
The root mean square error (RMSE) systematically represents the error that is
similar to all points in the data. Accurate models have low systematic RMSE. The
advantage of this measure of skill is that it retains the unit of the forecast variable. It is
given by the following formula, where O and F are observed and forecast fields:
( )∑=
−=N
1i
2ii FO
N1RMSE
RMSE is not a perfect tool and weaknesses include favouring model forecasts that
underestimate variability (Brier and Allan 1951).
b. Contingency Tables and Skill Scores
Contingency tables are useful in assessing the goodness of forecasts that give us four
possibilities: a “hit” (forecast above normal matches with the observed above normal),
a “false alarm” (forecast above normal (YES), but observed below normal (NO)), a
“miss” (forecast below normal (NO), but observed above normal (YES)), and a
“correct rejection” (forecast below normal (NO) matches with the observed below
normal (NO)). This scenario may be represented in tabular form as shown below:
6
Forecast ‘Yes’ Forecast ‘No’
Observed ‘Yes’ a b
Observed ‘No’ c d
The Hit Rate (HR) is a ratio of the number of hits to the number of events. But
the hit rate tells us at least as much about how often events are forecast as it does
about how good the forecasts are. It is given by:
baaHR+
=
The Proportion Correct (PC) is the ratio of the number of correct forecasts to
the number of forecasts. But the proportion correct (or hit score) tells us at least as
much about how often events occur as it does about how good the forecasts are. It is
given by:
dcbadaPC+++
+=
The Frequency Bias Index (FBI) measures the event frequency with no regard
for the forecast accuracy. Its value is 100% for a perfect forecast, and it is larger
(smaller) than 100% if the system is over forecasting (under forecasting). It is given
by:
cabaFBI
+
+=
The Equitable Threat Score (ETS) is a modified version of the threat score
rendered equitable by taking away the random forecast. Therefore, a chance forecast
will score 0, as will a constant forecast. A perfect forecast will have an ETS equal to
100%. It is given by:
)()(aRcba
aRaETS−++
−= , where the random forecast is
dcbacabaaR
+++
++=
))(()(
The True Skill Statistic (TSS) or Hansen-Kuipers discriminant, like in the
ETS, give the random and constant forecasts a zero score, while a higher score is
obtained if a rare event is forecast correctly. The TSS can also be written as the
7
probability of detection [a/(a + c)] minus the probability of false detection [b/(b +d)].
It is given by:
))(( dbcabcadTSS++
−=
The Heidke skill score (HSS) is a popular verification statistic, often
associated with chance as the standard of comparison. It is the ratio of the difference
between the number of correct forecasts and the number expected to be correct based
on chance, persistence, climatology, etc.) to the difference between the total number of
forecasts and the number expected to be correct based on chance. The HSS is given
by:
)()()(
aR-dcbaaR-d)(a= HSS
++++ where
)())(())(()(
dcbadbdc+caba= aR
+++++++
The Two-Alternative Forced Choice Test (2AFC) test is a test to correctly
identify which of two options has a characteristic of interest. A score of 100%
indicates a perfect set of answers. A score of 0% indicates a perfectly bad set of
answers. A score of 50% would be expected by guessing the answer every time. A
score of more than 50% suggests we have some skill in answering the questions. The
2AFC is given by:
( )))(( dcbabdacadP 2
1
AFC2 ++
+−=
4. RSM Performance over Kenya
a. RSM Temperature Verification over Kenya for OND Season
Table 1-1a: Correlation Coefficients (r) , Variance (r2) and RMSE for RSM OND Rainfall Output over Kenya
OND Rainfall Verification for Dodomar=0.3, r2=0.07
-2.0
-1.0
0.0
1.0
2.0
3.0
4.0
1970
19
71 19
72 19
73 19
74 19
75 19
76 19
77 19
78 19
79 19
80 19
81 19
82 19
83 19
84 19
85 19
86 19
87 19
88 19
89 19
90 19
91 19
92 19
93 19
94 19
95 19
96 19
97 19
98 19
99
Year
Rai
nfal
l Ano
mal
y
Obs_ONDRSM_OND
Figure 17: Inter-annual rainfall variability over Dodoma, Tanzania during the OND rainfall season. RSM-Simulated values are also shown.
21
OND Rainfall Verification for Iringar=0.5, r2=0.24
-3.0
-2.0
-1.0
0.0
1.0
2.0
3.0
4.0
1970
19
71 19
72 19
73 19
74 19
75 19
76 19
77 19
78 19
79 19
80 19
81 19
82 19
83 19
84 19
85 19
86 19
87 19
88 19
89 19
90 19
91 19
92 19
93 19
94 19
95 19
96 19
97 19
98 19
99
Year
Rai
nfal
l Ano
mal
y
Obs_ONDRSM_OND
Figure 18: Inter-annual rainfall variability over Iringa, Tanzania during the OND rainfall season. RSM-Simulated values are also shown.
OND Rainfall Verification for DIA
r=0.4, r2=0.16
-2.0
-1.5
-1.0
-0.5
0.0
0.5
1.0
1.5
2.0
2.5
3.0
1970
19
71 19
72 19
73 19
74 19
75 19
76 19
77 19
78 19
79 19
80 19
81 19
82 19
83 19
84 19
85 19
86 19
87 19
88 19
89 19
90 19
91 19
92 19
93 19
94 19
95 19
96 19
97 19
98 19
99
Year
Rai
nfal
l Ano
mal
y
Obs_ONDRSM_OND
Figure 19: Inter-annual rainfall variability over Dar es Salaam International Airport, Tanzania during the OND rainfall season. RSM-Simulated values are also shown.
22
OND Rainfall Verification for Mtwarar=0.2, r2=0.03
-2.0
-1.5
-1.0
-0.5
0.0
0.5
1.0
1.5
2.0
2.5
3.0
3.5
1970
19
71 19
72 19
73 19
74 19
75 19
76 19
77 19
78 19
79 19
80 19
81 19
82 19
83 19
84 19
85 19
86 19
87 19
88 19
89 19
90 19
91 19
92 19
93 19
94 19
95 19
96 19
97 19
98 19
99
Year
Rai
nfal
l Ano
mal
y
Obs_ONDRSM_OND
Figure 20: Inter-annual rainfall variability over Mtwara, Tanzania during the OND rainfall season. RSM-Simulated values are also shown.
b. RSM Rainfall Verification over Tanzania for MAM Season
Table 2-2a: Correlation Coefficients (r) and Variance (r 2) for RSM and Observed MAM Rainfall Output over Tanzania
Figure 57: Inter-annual rainfall variability over Abu Naama, Sudan during the JJAS rainfall season. RSM-Simulated values are also shown.
9. Dynamical Predictability over the Greater Horn of Africa Using RSM
a. Degree of Predictability over Greater Horn of Africa
The goal of seasonal climate forecasting is to provide advance knowledge about the
expected conditions of the atmosphere to interested users. Predictability over the
51
GHA region is classified as short range (1-3 days), medium range (4-7 days, but may
extend to 8-10 days), and long range (beyond two weeks). The skill of forecasts
exceeding two weeks is enhanced by combining statistical and dynamical techniques,
including assumption of persistence of synoptic systems.
Does the RSM dynamical model capture the seasonality of weather well? By
and large, the pressure, and hence wind fields are fairly well simulated by dynamical
models.
The accuracy of simulating the seasonal/annual cycle of convection and
precipitation over the region is still poor. The skill of predictability is better for the
short-term range, and is best within the first 24 to 48 hours, and decreases
drammatically thereafter.
RSM simulates fairly well the general pattern of the seasonal migration of the
ITCZ over the GHA. However, the accuracy of dynamically downscaled prediction
products still poses a challenge over the area. The seasonal predictability over
Equatorial Eastern Africa is dramatically better during the short rainfall season
(October-December), when the weather systems and oceanic controls are steadier and
better organized than the long rainfall season (March-May). Unfortunately, the
predictability using RSM over the northern parts of the GHA in JJA season, when the
region experiences most rainfall, is still wanting.
The actual convective processes that govern “tropical cyclogenesis” and
rainfall generation in the region have not been addressed in dynamical modelling over
the region. This can be determined by the sensitivity analysis of cloud simulation
modules in the dynamical models applied.
b. The Challenge of Limitation to Predictability
The interaction between mesoscale forcing with large scale circulation is one of the
most important considerations in the widespread local convection over the Greater
Horn of Africa region.
Errors arise in the prediction of dynamical and physical processes, including
tropical convection, however, due to a variety of reasons including inadequate
horizontal resolution, improper cascading of energy from one scale to another,
unsuitable parameterization schemes, et cetera. The cumulative impact of these
modelling uncertainties on the predicted quantities may be significant. This is
aggravated by a lack of understanding of the roles of complex nonlinear interactions
and feedbacks between systems of different scales on the temporal evolution of
52
convection. Most cumulus convection modules in dynamical models, being ad hoc
and presuming linearity, fail to capture the diurnal and seasonal cycles of convection
accurately.
The subject of limitation to predictability in the Greater Horn of Africa region
needs to be addressed so as to meet one of the main goals of dynamical forecasting,
which is to improve the predictability of monthly and seasonal climate events on
regional to global scales to meet the sundry needs of users.
References
Anyah, R.O., 2005: Modeling the Variability of the Climate System over Lake Victoria Basin. PhD Thesis, North Carolina State University, USA.
Asnani G.C. and J.H Kinuthia, 1979: Diurnal variation of precipitation in East Africa. Kenya Meteorological Department. Research Report, No. 8/79, 1-58.
Asnani, G.C., 1993: Tropical Meteorology, Volume I and II, Sindh Colony, Aundh, Pune, India.
Camberlin, P. and N. Philippon, 2002: The East African March-May rainy season: Associated atmospheric dynamics and predictability over the 1968-97 period. J. Climate, 15, 1002-1019.
Indeje, M., F.H.M. Semazzi, L. Xie, L.J. Ogallo, 2001: Mechanistic Model Simulations of the East African Climate Using NCAR Regional Climate Model: Influence of Large-Scale orography on the Turkana Low-Level Jet. J. Climate, 14, 2710-2724.
Ininda, J. M, 1998: Simulation of the impact of sea surface temperature anomalies on the short rains over East Africa. J. African Meteorology Society, 3, 1, 127-140.
Mukabana, J.R., 1992: Numerical simulation of the influence of large scale monsoon flow on the diurnal weather patterns over Kenya. PhD Thesis, University of Nairobi.
Mutai, C.C. and M.N. Ward, 2000: East African rainfall and the tropical circulation/convection on intraseasonal to interannual timescales. J. Climate, 13, 3915-3939.
Mutai, C.C. M.N. Ward, and A.W. Coleman, 1988: Towards the prediction of the East Africa short rains based on sea-surface temperature-atmosphere coupling. Int. J. Climatol., 18, 975-997.
Mutemi, J. N, 2003: Climate anomalies over eastern Africa associated with various ENSO evolution phases. Ph.D. Thesis, University of Nairobi, Kenya.
Okeyo, A.E. 1987: Towards the development of a forecasting numerical model for Kenya. Ph.D. Thesis. University of Nairobi, Kenya.
Opijah, F.J., 2000: Numerical Simulation of the Impact of Urbanization on the Microclimate over Nairobi Area. PhD Thesis. University of Nairobi, Kenya.
Sakwa, V.N, 2004: Assessment of the Skill of the High Resolution Regional Model in the Simulation of Airflow and Rainfall over East Africa. MSc Dissertation. University of Nairobi, Kenya.