Content uploaded by Steve Twomlow
Author content
All content in this area was uploaded by Steve Twomlow on Nov 09, 2017
Content may be subject to copyright.
Start, end and dry spells of the growing season in semi-arid southern Zimbabwe
W. Mupangwa 1*, S. Walker 2
,
S
. Twomlow 3
1*ICRISAT-Bulawayo, Matopos Research Station, P O Box 776, Bulawayo, Zimbabwe
2Department of Soil, Crop and Climate Sciences, University of Free State, Bloemfontein,
South Africa
3United Nations Environment Programme, P O Box 30552 (00100), Nairobi, Kenya
Abstract
Smallholder agriculture in semi-arid Zimbabwe is dependant on seasonal characteristics
of rainfall. The determination of start, end and length of the growing season, and the
pattern of dry spells during the season is useful information for the planning of land
preparation and planting activities. This study was designed to assess whether there has
been any changes in the start, end and length of growing season and the pattern of 14 and
21 day dry spells during the season. Daily rainfall data were collected from five
meteorological stations located in southern Zimbabwe. Results indicated that no
significant changes in the start, end and subsequent length of growing season have
occurred over the past 50–74 years. The number of wet days per growing season has not
changed significantly either. There are high chances of 14 and 21 day dry spells during
the peak rainfall months. A strong correlation exists between start and end of the growing
season in agroecological region V. We conclude that growing seasons have not changed
significantly over the past 50–74 years in southern Zimbabwe. There is scope in
exploring rainwater management technologies to reduce the impact of dry spells in
rainfed cropping systems.
1 * Corresponding author: w.mupangwa@cgiar.or or Mupangwa@yahoo.com (Walter Mupangwa)..
1
Key words: Dry spell, end of season, length of growing season, start of season, semi-
arid, wet days
Introduction
Semi-arid southern Zimbabwe experiences frequent droughts and dry spells during the
growing season, making rainfed cropping risky (Cooper et al., 2008). In some years the
rains start early whereas in others they arrive late. An abrupt end of the growing season in
semi-arid parts of Zimbabwe has been experienced in some years (Mupangwa, 2008).
This annual variability makes the selection of crop types and varieties, and planning of
planting dates critical, yet also difficult, for successful cropping in rainfed systems
(Hussein, 1987; Kinsey et al., 1998; Raes et al., 2004). Crop yields are often reduced
significantly due to the late start and early cessation of the growing season. This is further
complicated by the occurrence of long dry spells during the January to February period
when most crops are in their vegetative and reproductive growth stages. Increases in dry
spell lengths and reductions in wet day frequencies have been reported in Malawi,
Zambia and Zimbabwe (Tadross et al., 2007). Effective rainfall for planting has been
arriving late over much of southern Africa (Tadross et al., 2007). In southern Zimbabwe,
as in other parts of sub-Saharan Africa, it is now very common for drought or long dry
spells to occur during the growing season (Usman and Reason, 2004).
The existence of relationships between start, end and length of growing season, and
number of wet days per growing season is critical for planning farming activities before
the start and during the season (Mugalavai et al., 2008). Studies conducted in semi-arid
2
parts of West Africa indicated that there is a significant relationship between the start of
rains and the length of the growing season (Sivakumar, 1988). The length of the growing
season is more sensitive to the start of the rains than to the cessation (Oladipo and Kyari,
1993). Analysis of rainfall data from the more humid northern Zimbabwe indicated that
the length of growing season increases with earlier start of the rains (Chiduza, 1995). A
strong (R2 = 0.76) relationship between start and length of the growing season has been
reported in some parts of semi-arid southern Africa (Kanemasu et al., 1990).
This study was designed to assess changes in the start, end and length of the growing
season using historical daily rainfall data derived from five meteorological stations
located in semi-arid Mzingwane catchment of southern Zimbabwe. The study also
explored the patterns of wet days, 14 and 21 day dry spells during the growing season.
The relationships between these characteristics of the growing season were determined.
Materials and Methods
Site description
Historical rainfall data were collected from five meteorological stations located in
Matebeleland South province of southern Zimbabwe (Table 1) which is part of the
Limpopo River Basin (Fig. 1). The Limpopo River Basin lies in southern Africa between
20 and 26oS, and 25 and 35oE (FAO, 2004). Total annual rainfall (based on the July–June
calendar) varies from 584 mm at Bulawayo (Mupangwa, 2008) to less than 400 mm in
the south eastern lowveld (Thuli and Beitbridge) in the Zimbabwe part of the Limpopo
basin (FAO, 2004; Unganai, 1996). The rainy season spans from October to April with
3
the peak rainfall period occurring between December and February. Periodic in-season
dry spells occur frequently during the January to February period impacting negatively on
crop production in southern Zimbabwe. The average daily maximum temperatures vary
from 30–34oC during summer to 22–26oC in winter. Minimum temperatures average 18–
22oC during summer and 5–10oC in winter (FAO, 2004). Evaporation in the Limpopo
river basin varies from 1 600 mm/year to more than 2 600 mm/year and is highest during
the rainfall season (FAO, 2004).
Table 1. Geographical description of meteorological stations and rainfall database of the
five stations used in the analyses
Station Latitude Longitude Data
period Duration
of data set Agroecological
region
Bulawayo -20.22 28.62 1930-2001 71 IV
Matopos -20.38 28.50 1939-2008 69 IV
Mbalabala -20.35 28.95 1931-1994 64 IV
Filabusi -20.55 29.28 1921-1995 74 IV
Beitbridge -22.22 30.00 1951-2001 50 V
27 °E 30 ° 33 °
21 °S
24 °
Indian
Ocean
Mwenezi
Limpop o
Motloutse
Shashe
Tuli
Bubye
Changane
Olifants
Great
Letaba
Middle
Letaba
Ngotwane
Marico
Crocodile
Mokolo
Lapha lala
Mogalakwena
Sand
Blyde
Reit
Wilge
Elands
Pienaars
Levuvhu
Lotsan e
300
400
500
600
700
800
900
1000
Mean Annual
R ain fa ll (m m )
1250
N
Selati
Umzingwani
Little
Olifants
Steelpoort
Shingwedzi
0 50 100 150 200 km
ZIMBABWE
MOZAMBIQUE
MOZAMBIQUE
BOTSW ANA
SOUTH
AFRICA
Mzingwane Catchment
Figure 1. The Limpopo River Basin showing the distribution of mean annual rainfall over
the basin
4
Dry and wet days
Meteorologically speaking a wet day definition of ‘any day with more than 0.85 mm
accumulated in 1 or 2 days’ might suffice. However, for crop production purposes in a
region experiencing daily pan evaporation of 5–8 mm (Woltering, 2005) rainfall of 0.85
mm has very limited influence on crop growth. Stern et al. (2003) suggest that for other
purposes a threshold of 4.95 mm can be used to define a wet day. For the purposes of our
study in semi-arid southern Zimbabwe the following definitions were used for dry and
wet days:
• Dry day: Any day that accumulates less than 5 mm of rain
• Wet day: Any day that accumulates 5 mm or more
Start, end and length of growing season
The start, end and length of the growing season were defined as:
• Start: the first day after 1 October when the rainfall accumulated over 1 or 2 days
is at least 20 mm and not followed by a period of more than 10 consecutive dry
days in the following 30 days (Stern et al., 2003)
• End: the last day before 30 June that accumulates 10 mm or more rainfall
• Length of growing season: This was calculated by subtracting the date of the
beginning from the date of ending of the growing season
The condition of having no dry spell of more than 10 days after start of growing season
eliminates the possibility of a false start of the season. A period of 30 days is the average
length for the initial growth stage of most crops (Allen et al., 1998). Most crops would
5
have emerged and be well established after 30 days. The cut off point for end of season
catered for late maturing or late planted crops. After 1 June temperature normally drops
quite significantly (FAO, 2004) and crop growth rate is slowed down (Evans, 1979;
Sundararaj and Thulasidas, 1980). Given the above definitions Instat Statistical
programme (Version 3.33) (Stern et al., 2003) was used to analyse the daily rainfall data
for start and end of season, and length of the growing season using the July to June
calendar.
The F- and t-tests, to confirm significance of change in start and end of growing season,
length of growing season and number of wet days per growing season, were conducted
using Genstat Discovery Edition 3 (www.vsni.co.uk). Regression analysis was conducted
to determine the relationship between (a) start and length of growing season, (b) start and
end of growing season, (c) start and number of wet days per growing season, and (d)
number of wet days and length of growing season. A t-test was applied to check whether
slopes of the regressions were significantly different from zero.
Dry spell analysis
Daily rainfall data for each station were fitted to the simple Markov chain model as
outlined by Stern et al. (1982; 2003). The Markov chain model was run so that it will
give the probability of getting 14 and 21 day dry spells within 30 days following a wet
day using the July to June calendar. The analyses were performed using Instat Statistical
Programme (Version 3.33, http://www.reading.ac.uk/ssc/software/instat/climatic.pdf.)
(Stern et al., 2003).
6
Results and Discussion
Start and end of growing season
There has been no significant (P > 0.05) change in the start and end of growing seasons at
each station across the Bulawayo to Beitbridge transect (Table 2, Figs. 2 and 3). The
variability of start and end of the growing season is quite similar at all stations along the
Bulawayo to Beitbridge transect. The growing season generally starts earlier at Filabusi
and ends earlier at Beitbridge (Table 2). In Bulawayo the growing season starts on 8
December and ends on 4 April (Table 2) giving an average season length of 117 days. At
Matopos the growing season usually starts on 2 December and ends on 29 March. At
Mbalabala the growing season starts on 3 December and ends on 1 April. Growing
season starting as late as 18 March at Mbalabala (Table 3) can be attributed to the criteria
of start of season of 20 mm of rain in one or two days used in our study. At Filabusi the
growing season starts on 28 November and ends on 3 April station. At Beitbridge which
was the driest station studied, the season starts on 7 December.
Table 2. Median dates for the start and end of the growing season based on daily rainfall
data obtained from five meteorological stations in southern Zimbabwe
Station Median start
date Standard
deviation
(days)
Median end
date Standard
deviation
(days)
Bulawayo 8 December 31 4 April 32
Matopos 2 December 30 29 March 31
Mbalabala 3 December 26 1 April 27
Filabusi 28 November 26 3 April 30
Beitbridge 7 December 31 25 March 29
The most abrupt end of season at Bulawayo was recorded in 1964/65 when the season
ended on 15 January (day 200) (Table 3). At Beitbridge, which lies in the most arid part
7
of the transect, the season ends earlier than other stations along the transect. This is
consistent with results from Aviad et al. (2004) which showed that as aridity increases
the season starts late and ends early. The most abrupt end of season at Beitbridge was
recorded on 20 January. The most delayed end of season was in 1958/59 and occurred on
13 June. The abrupt end of the growing seasons observed in the long-term analysis agrees
with observations made in semi-arid southern Zimbabwe during the 2007/08 growing
season (Mupangwa, 2008). The 2007/08 growing season ended on 24 January 2008 in
semi-arid districts of southern Zimbabwe.
Table 3. Dates for the earliest and most delayed start and end of the growing season along
the Bulawayo to Beitbridge transect
Station Earliest start Most delayed
start Earliest end Most delayed
end
Bulawayo 28 October 3 March 15 January 16 June
Matopos 21 October 23 February 13 December 1 June
Mbalabala 30 October 18 March 21 January 16 May
Filabusi 10 October 14 February 20 January 13 June
Beitbridge 10 October 14 February 20 January 13 June
Aviad et al. (2004) from the Mediterranean region reported that as aridity increases the
length of growing season is short. In our study smallholder farmers in Beitbridge face the
prospects of a shorter growing season compared to those in NR IV as it only began on 7
December and ended on 25 March giving a length of only 108 days. This is further
complicated by the fact that start of growing season is also more variable in Beitbridge
than its end (Table 2). Farming practices that allow timely preparation of the land and
planting are more critical in the Beitbridge area so that farmers can make effective use of
rainfall received during the November to December period.
8
Bulawayo
0
50
100
150
200
250
300
350
400
21/22
25/26
31/32
35/36
39/40
43/44
47/48
51/52
55/56
59/60
63/64
67/68
71/72
75/76
79/80
83/84
87/88
91/92
95/96
99/00
03/04
07/08
Season
Time (days after 1 July
)
5 season moving average 5 season moving average Start End
a
Matopos
0
50
100
150
200
250
300
350
400
21/22
25/26
31/32
35/36
39/40
43/44
47/48
51/52
55/56
59/60
63/64
67/68
71/72
75/76
79/80
83/84
87/88
91/92
95/96
99/00
03/04
07/08
Season
Time (days from 1 July
)
5 season moving average 5 season moving average Start End
Mbalabala
0
50
100
150
200
250
300
350
400
21/22
25/26
31/32
35/36
39/40
43/44
47/48
51/52
55/56
59/60
63/64
67/68
71/72
75/76
79/80
83/84
87/88
91/92
95/96
99/00
03/04
07/08
Season
Time (days after 1 July)
5 season moving average 5 season moving average Start End
b
c
Figure 2. Start and end of growing seasons derived from daily rainfall data for Bulawayo
(a), Matopos (b) and Mbalabala (c) Meteorological stations
9
Filabusi
0
50
100
150
200
250
300
350
400
21/22
25/26
31/32
35/36
39/40
43/44
47/48
51/52
55/56
59/60
63/64
67/68
71/72
75/76
79/80
83/84
87/88
91/92
95/96
99/00
03/04
07/08
Season
Time (days after 1 July)
5 season moving average
5 season moving average
Start
End
Beitbridge
0
50
100
150
200
250
300
350
400
21/22
25/26
31/32
35/36
39/40
43/44
47/48
51/52
55/56
59/60
63/64
67/68
71/72
75/76
79/80
83/84
87/88
91/92
95/96
99/00
03/04
07/08
Season
Time (days afetr 1 July)
5 season moving average
5 season moving average
Start
End
d
e
Figure 3. Start and end of growing seasons derived from daily rainfall data for Filabusi
(d) and Beitbridge (e) meteorological stations
The lack of significant long-term changes suggests that the characteristics of the growing
season are influenced by other factors in addition to total rainfall and date of start of the
rains. In Zimbabwe, Oosterhout (1996) reported that years that had the highest rainfall
did not correspond with years having the highest crop yields. Distribution and reliability
of rainfall during the growing season has a stronger influence on the characteristics of the
growing period than total rainfall (Dennett, 1987; Adiku et al., 1997). High spatial and
temporal rainfall variability during the season increases the risk of intra seasonal dry
spells and droughts in semi-arid areas (Rockström and Falkenmark, 2000). Long-term
simulation results (Mupangwa, 2008) indicated that there is more than a 20% chance of
10
getting no yield in the conventional and CA systems under semi-arid conditions of
southern Zimbabwe. Therefore, other indices to be able to represent this distribution and
variability need to be developed.
Length of growing season
The length of growing season decreased as one moves from Bulawayo to Beitbridge
through Matopos, Mbalabala and Filabusi (Figs. 4 and 5). For the period reviewed at each
station, the growing season averaged 111 days at Bulawayo, 110 days at Matopos, 112
days at Mbalabala, 122 days at Filabusi and 100 days at Beitbridge. Based on these
results sorghum varieties such as SV 2, SV 4 and Macia with 110–127 days to maturity
(Mgonja et al., 2005) can be successfully grown around the Bulawayo, Matopos,
Mbalabala and Filabusi stations. Pearl millet varieties such as PMV 2 and PMV 3 which
require 80–90 days to reach maturity (Monyo, 2002) can be grown along the Bulawayo to
Beitbridge transect. The prospect of having a crop reach maturity gives smallholder
farmers in southern Zimbabwe an incentive to plant as early as possible.
The longest growing season recorded across the five stations was 224 days at Matopos in
2001/2002 which is more than double the average length. The shortest growing season
observed across the Bulawayo to Beitbridge transect was 38 days recorded at Filabusi
and Beitbridge. The longest growing season at Bulawayo had 215 days and was recorded
in 1939/40. The 1981/82 season was the shortest growing season with only 41 days
which coincided with the lowest number of wet days (4). At Matopos the shortest
growing season had 39 days recorded in 1979/80. In 1954/55 Mbalabala recorded the
11
longest season which had 173 days which was also the season with the highest number
(48) of wet days. The shortest growing season at the same station was 46 days recorded
in 1976/77. Filabusi recorded the longest growing season of 197 days in 1958/59 while
the shortest season was 38 days recorded in 1946/47 which was the year with annual
rainfall of 250 mm. At Beitbridge the longest growing season was 162 days recorded in
1999/00 which also recorded the highest number of wet days (31) and the highest annual
rainfall (1177 mm) due to the cyclone Eline pushing in over the African landmass and
causing severe flooding. The shortest season of 38 days was recorded in 1972/73 and
1973/74.
Results from our study showed an average length of the growing season of 111 days
across the five stations. The 111 days length of growing season is much higher than 96
days reported by Morse (1996) for Zimbabwe’s NRs II, IV and V. The differences in
length of growing season could be ascribed to differences in the criteria used in defining
the start and end of growing season. Tadross et al. (2007) reported that a growing season
of 90–120 days can be experienced in southern African countries such as Malawi and
Zambia. The above analysis confirms that there is no general rule of thumb that the
season length is related to the amount of rainfall received but can have a relationship with
the number of wet days or effective rainfall events. So if this southern part of Zimbabwe
has an average season length of only 111 days, agronomists need to use this information
in selecting suitable cultivars for smallholder farmers. The ultra short cultivars could be
an option and this will help farmers to be able to plant early and produce a crop with less
exposure to dry spells.
12
Bulawayo
0
50
100
150
200
250
21/22
25/26
31/32
35/36
39/40
43/44
47/48
51/52
55/56
59/60
63/64
67/68
71/72
75/76
79/80
83/84
87/88
91/92
95/96
99/00
03/04
07/08
Season
Length of season (days
)
Growin
g
season 5 season movin
g
avera
g
e
Matopos
0
50
100
150
200
250
21/22
25/26
31/32
35/36
39/40
43/44
47/48
51/52
55/56
59/60
63/64
67/68
71/72
75/76
79/80
83/84
87/88
91/92
95/96
99/00
03/04
07/08
Season
Length of season (days)
Growing season 5 season moving average
Mbalabala
0
50
100
150
200
250
21/22
25/26
29/30
33/34
37/38
41/42
45/46
49/50
53/54
57/58
61/62
66/67
70/71
74/75
78/79
82/83
86/87
90/91
94/95
98/99
02/03
06/07
Season
Length of season (days
)
Growin
g
season 5 season movin
g
avera
g
e
a
b
c
Figure 4. Length of the growing season based on daily rainfall data obtained from
Bulawayo (a), Matopos (b) and Mbalabala (c) meteorological stations
13
Filabusi
0
50
100
150
200
250
21/22
25/26
31/32
35/36
39/40
43/44
47/48
51/52
55/56
59/60
63/64
67/68
71/72
75/76
79/80
83/84
87/88
91/92
95/96
99/00
03/04
07/08
Season
Length of season (days
)
Growing season 5 season moving averag
e
d
Beitbridge
0
50
100
150
200
250
21/22
25/26
31/32
35/36
39/40
43/44
47/48
51/52
55/56
59/60
63/64
67/68
71/72
75/76
79/80
83/84
87/88
91/92
95/96
99/00
03/04
07/08
Season
Length of season (days)
Growing season 5 season moving average
e
Figure 5. Length of the growing season based on daily rainfall data obtained from
Filabusi (d) and Beitbridge (e) meteorological stations
Wet days per growing season
The changes in number of wet days per growing season based on daily rainfall data
derived from the five meteorological stations in southern Zimbabwe are shown in Figures
6 and 7. There was a negligible difference in the number of wet days per season as one
moves from Bulawayo to Filabusi station through Matopos and Mbalabala. All four
stations are located in NR IV. Our findings are inconsistent with results reported by
Hudson and Jones (2002) who stated that there has been a general decrease in the number
of wet days per year in southern Africa. Hudson and Jones (2002) defined a wet day as a
14
day receiving more than 0.2 mm of rain whereas in our study we defined a day with 5
mm or more rainfall as wet. In addition, our study focussed on number of wet days per
growing season and not the whole year. As expected Beitbridge which lies in NR V, had
the least (P = 0.01) number of wet days per growing season (Fig. 7).
Bulawayo
0
5
10
15
20
25
30
35
40
45
50
21/22
25/26
31/32
35/36
39/40
43/44
47/48
51/52
55/56
59/60
63/64
67/68
71/72
75/76
79/80
83/84
87/88
91/92
95/96
99/00
2003/04
2007/08
Season
Number of wet days
5season movin
g
avera
g
eWet da
y
s
Matopos
0
5
10
15
20
25
30
35
40
45
50
21/22
25/26
31/32
35/36
39/40
43/44
47/48
51/52
55/56
59/60
63/64
67/68
71/72
75/76
79/80
83/84
87/88
91/92
95/96
99/00
03/04
07/08
Season
Number of wet days
5season moving average Wet days
a
b
Figure 6. Number of wet days per season based on daily rainfall data obtained from
Bulawayo (a) and Matopos (b) meteorological stations
15
Mbalabala
0
5
10
15
20
25
30
35
40
45
50
21/22
25/26
31/32
35/36
39/40
43/44
47/48
51/52
55/56
59/60
63/64
67/68
71/72
75/76
79/80
83/84
87/88
91/92
95/96
99/00
03/04
07/08
Season
Number of wet days
5 season movin
g
avera
g
eWet da
y
s
Filabusi
0
5
10
15
20
25
30
35
40
45
50
21/22
25/26
31/32
35/36
39/40
43/44
47/48
51/52
55/56
59/60
63/64
67/68
71/72
75/76
79/80
83/84
87/88
91/92
95/96
99/00
03/04
07/08
Season
Number of wet days
5season movin
g
avera
g
e
Wet da
y
s
Beitbridge
0
5
10
15
20
25
30
35
40
45
50
21/22
25/26
31/32
35/36
39/40
43/44
47/48
51/52
55/56
59/60
63/64
67/68
71/72
75/76
79/80
83/84
87/88
91/92
95/96
99/00
03/04
07/08
Season
Number of wet days
5season movin
g
avera
g
eWet da
y
s
c
d
e
Figure 7. Number of wet days per season based on daily rainfall data obtained from
Mbalabala (c), Filabusi (d) and Beitbridge (e) meteorological stations
16
The average number of wet days per growing season recorded at Bulawayo was 21,
Matopos was 22, Mbalabala was 24, Filabusi was 22 and Beitbridge was 12. At
Bulawayo station the lowest number of wet days per season recorded was four in 1946/47
and 1981/82. The 1981/82 growing season was recorded as the driest in other southern
African countries (Cook et al., 2004). Forty-two wet days were recorded in 1977/78 was
the highest number observed at Bulawayo (Fig. 6). Matopos recorded seven and 45 wet
days in 1990/91 and 1954/55 seasons. Mbalabala recorded eight and 48 wet days in
1983/84 and 1954/55 seasons (Fig. 7). Filabusi station recorded eight and 46 wet days per
growing season in 1946/47 and 1954/55 seasons. Three and 31 wet days per growing
season were recorded at Beitbridge station in 1964/65 and 1999/00 seasons. Using the
October–September hydrological year and definitions of a wet day having more than 10,
20 and 30 mm rainfall, Love et al. (2008) assessed the trend in number of wet days per
year in southern Zimbabwe. Findings from their study revealed a decline in number of
days with more than 10, 20 and 30 mm of rainfall per year in southern Zimbabwe.
However, this is not apparent from the current analysis with a wet day defined as above 5
mm.
At Bulawayo and Matopos the driest decade (1980–1990) had the lowest number of wet
days per season. At Bulawayo and Matopos the return period of seasons with high
rainfall is 18 to 22 years. The 1961 to 1978 period had the least number of wet days per
season at Mbalabala (Fig. 7) which coincides with the period of lowest annual rainfall.
Return period for high number of wet days per season was 21–26 years at Mbalabala and
17
Filabusi. Further south, Beitbridge appears to receive relatively wet growing seasons after
more than 20 years although the record is only 50 years long.
Dry spells during the growing season
The peak rainfall period during the growing season occurs between December and
February in southern Zimbabwe. The likelihood of getting 14 and 21 day dry spells
during any time of the year is given in Figures 8 and 9. Based on the data set reviewed,
Filabusi has the highest chance of getting 14 and 21 day dry spells compared to the other
four stations during the peak rainfall period (Fig. 9) suggesting that rainfed crop
production is more risk at Filabusi. Mbalabala has the least chance of getting a 21 day dry
spell during the peak rainfall period from 120 to 240 days after 1 July. The probability of
getting a 14 or 21 day dry spell decreases as one move from Bulawayo to Mbalabala
through Matopos. Mbalabala experiences slightly more rainy days per growing season
than the other four stations (Figs. 6 and 7). There is a 27 to 40% chance of experiencing a
21 day dry spell at Bulawayo between 10 November (day 133) and 8 February (day 223).
During the same period there is a 66 to 79% chance of getting a 14 day dry spell at
Bulawayo. At Matopos there is a 18 to 33% chance of a 14 day dry spell occurring
between 10 November and 8 February. During the same period 21 day dry spells have a 3
to 7% chance of occurring at Matopos.
The probability of getting a 14 day dry spell at Mbalabala between 10 November and 8
February is 14 to 27%. There is a 2 to 5% chance of getting 21 continuous dry days
during the same period. Fourteen day dry spells are a common feature at Filabusi with a
18
91% chance of occurring between 10 November and 8 February. During the same period
there is a 60 to 80% probability of a 21 day dry spell occurring. Further down the transect
Beitbridge has a 48 to 69% chance of experiencing a two week dry spell. A 14 to 30%
chance exists for a 21 day dry spell to occur between 10 November and 8 February at
Beitbridge.
The decrease in probability of getting 14 and 21 day dry spells coincides with the peak
rainfall period. The peak rainfall period occurs from December to February (Unganai,
1996). Figures 8 and 9 reveal that rainfall is more reliable at Matopos and Mbalabala
during the December to February peak rainfall period. At Beitbridge there is a rapid
increase in chances for experiencing 14 and 21 day dry spells after receiving rain. This
implies that the probability of reduced crop yields or complete crop failure due to soil
water deficits is also high at Beitbridge. The probability of dry spells also increases at
Bulawayo, Matopos, and Mbalabala after 8 February. This coincides with the flowering
and grain filling stages of most cereals grown in semi-arid smallholder systems. This
poses a major challenge in soil water management in semi-arid rainfed cropping systems.
19
Bulawayo
0.0
0.2
0.4
0.6
0.8
1.0
0 50 100 150 200 250 300 350 400
Time (days after 1 July)
Probabilit
y
14
21
Matopos
0.0
0.2
0.4
0.6
0.8
1.0
0 50 100 150 200 250 300 350 400
Time (days after 1 July)
Probabilit
y
14
21
Mbalabala
0.0
0.2
0.4
0.6
0.8
1.0
0 50 100 150 200 250 300 350 400
Time (days after 1 July)
Probabilit
y
14 21
a
b
c
Figure 8. Probability of getting 14 and 21 day spells within 30 days from a wet day based
on the fitted first order Markov chain probability values for Bulawayo (a) and Matopos
(b) meteorological stations in southern Zimbabwe
20
Filabusi
0.0
0.2
0.4
0.6
0.8
1.0
0 50 100 150 200 250 300 350 400
Time (days after 1 July)
Probabilit
y
14 21
Beitbridge
0.0
0.2
0.4
0.6
0.8
1.0
0 50 100 150 200 250 300 350 400
Time (days after 1 July)
Probabilit
y
14 21
d
e
Figure 9. Probability of getting 14 and 21 day spells within 30 days from a wet day based
on the fitted first order Markov chain probability values Mbalabala (c), Filabusi (d) and
Beitbridge (e) stations in southern Zimbabwe
Relationships of growing season characteristics
The correlation analysis indicated that there is no significant relationship between the
start and end of the growing season at Bulawayo and Filabusi stations. A stronger
relationship between start and end of growing season exists at Beitbridge than Matopos
and Mbalabala (Table 4). The gradient of start and end of growing season relationship is
steepest at Beitbridge station implying that delayed start of season is likely followed by
early cessation of the growing season. It is therefore more critical that smallholder
farmers in Beitbridge access seasonal weather forecast information in order to plan their
farming operations than those in Matopos and Mbalabala. There is a strong linear
21
relationship between the start and length of the growing season at all five stations. The
strength of the linear relationship decreases as one moves from Bulawayo (NR IV) to
Beitbridge (NR V) indicating more uncertainty in the length of the growing season in NR
V compared to NR IV.
Regression analysis revealed a significant (P < 0.001) inverse relationship between the
start and the number of wet days in a growing season (Table 4). A delayed start of the
growing season is likely to be accompanied by a reduction in the number of wet days in a
growing season implying a high risk of soil water deficits in the smallholder cropping
systems during the growing season. There is a significant linear relationship between the
number of wet days and the length of the growing season at four of the five stations. The
linear relationship suggests that more wet days translate into a longer growing season.
22
Table 4. Relationships of the different characteristics of the growing season in semi-arid
southern Zimbabwe
Station Relationship Regression
equation R2 value
Bulawayo Start and end Y = 251 + 0.16x 1.0
Start and length of
growing season Y = 251 – 0.84x 0.40
Start and number of
wet days per-season Y = 53 – 0.19x 0.38
Number of wet days
and length of
growing season
Y = 48.9 + 2.9x 0.46
Matopos Start and end Y = 223 + 0.30x 0.07
Start and length of
growing season Y = 223 – 0.70x 0.31
Start and number of
wet days per-season Y = 47.9 – 0.16x 0.25
Number of wet days
and length of
growing season
Y = 57.9 + 2.4x 0.35
Mbalabala Start and end Y = 217 + 0.35x 0.09
Start and length of
growing season Y = 217 – 0.65x 0.29
Start and number of
wet days per-season Y = 48.2 – 0.15x 0.16
Number of wet days
and length of
growing season
Y = 65.3 + 1.9x 0.36
Filabusi Start and end Y = 244 + 0.21x 0.02
Start and length of
growing season Y = 244 – 0.80x 0.33
Start and number of
wet days per-season Y = 43.0 – 0.14x 0.16
Number of wet days
and length of
growing season
Y = 79.4 + 1.9x 0.22
Beitbridge Start and end Y = 190 + 0.47x 0.23
Start and length of
growing season Y = 190 – 0.53x 0.27
Start and number of
wet days per-season Y = 26.6 – 0.09x 0.16
Number of wet days
and length of
growing season
Y = 66.6 + 2.85x 0.34
23
Conclusion and recommendations
Analysis of the characteristics of the growing season has demonstrated that there have
been insignificant changes in the start, end and length of growing season along the
Bulawayo to Beitbridge transect in southern Zimbabwe. The trend of the number of wet
days per growing season has been weak over the 50–75 years reviewed. Despite the lack
of significant changes in the characteristics of the growing season, delayed start, early
cessation and subsequent short growing seasons have been experienced during some
years over the 50–75 year period reviewed. The growing season ends much earlier in
Beitbridge (NR V) which is more arid compared to the stations in NR IV. This places
strain on smallholder farmers in Matebeleland South province and therefore it is
imperative to prepare land and plant on time if the recommended crop species and
varieties are to reach maturity. There are high chances of getting 14 and 21 day dry spells
during the peak rainfall period (January–February), further increasing the chances of
significant crop yield reduction and crop failure. There is scope in exploring rainwater
harvesting technologies in order to prolong the period of soil water availability in rainfed
smallholder cropping systems. Despite the lack of significant changes in the start, end
and length of the growing season there is need to establish crop planting windows for
major crops grown in the semi-arid areas of Zimbabwe.
Regression analysis indicate that at stations with the most delayed end of growing season
(Bulawayo and Filabusi), the start of growing season has no influence on the end of
season. As aridity increases from Bulawayo to Beitbridge, the strength of the start and
end of growing season relationship also increases. This suggests that delayed start of
24
growing season at Beitbridge is followed by early cessation of the growing season.
Similarly, early start could be translated into a longer growing season. Smallholder
farmers in southern Zimbabwe would need this information for deciding on crop types
and varieties, and dates for land preparation and planting. The inverse relationship
between start of season and number of wet days per growing season coupled with the
high chances of dry spells during the peak rainfall period suggest the possibility of soil
water deficits during the season. This further confirms the need to invest in rain and soil
water management technologies in the smallholder cropping systems of semi-arid
southern Zimbabwe.
Acknowledgements
This paper is a contribution to WaterNet Challenge Program Project 17 ‘Intergrated
Water Resource Management for Improved Rural Livelihoods: Managing risk, mitigating
drought and improving water productivity in the water scarce Limpopo Basin’. The
authors are grateful to the meteorological department for providing rainfall data.
References
Adiku, S.G.K., Dayananda, P.W.A., Rose, C.W. and Dowuona, G.N.N. 1997. An
analysis of the within-season rainfall characteristics and simulation of the daily
rainfall in two savanna zones in Ghana. Journal of Agricultural and Forest
Meteorology, 86: 51-62.
Allen, R.G., Pereira, L.S., Raes, D. and Smith, M. 1998. Crop evapotranspiration: a
guideline for computing crop water requirements. FAO Irrigation and Drainage
Paper No 56. FAO Water Resources, Development and Management Service,
Rome, Italy. 300pp.
Aviad, Y., Kutiel, H. and Lavee, H. 2004. Analysis of beginning, end and length of the
rainy season along a Mediterranean arid climate transect for geomorphic
purposes. Journal of Arid Environments, 59: 189-204.
25
Chiduza, C., 1995. Analysis of rainfall data and their implication on crop production: A
case of Northern Sebungwe. Zimbabwe Journal of Agricultural Research, 33
(2): 175-189.
Cook, C., Reason, C.J.C. and Hewitson, B.C. 2004. Wet and dry spells within particularly
wet and dry summers in the South African summer rainfall region. Climate
Research, 26: 17-31.
Cooper, P.J.M., Dimes, J., Rao, K.P.C., Shapiro, B. Shiferaw, B. and Twomlow, S. 2008.
Coping better with current climatic variability in the rain-fed farming systems of
sub-Saharan Africa: An essential first step in adapting to future climate change?
Agriculture, Ecosystems and Environment, 126: 24–35.
Dennett, M.D. 1987. Variation of rainfall – the background to soil and water management
in dryland regions. Soil Use and Management, 3(2): 47-51.
Evans, L.T. 1979. Crop Physiology. Blackie & Sons Pvt. Ltd., Bombay. 374pp.
FAO, 2004. Drought impact mitigation and prevention in the Limpopo River Basin. Land
and Water Discussion Paper 4. Food and Agriculture Organisation of the United
Nations, Rome. http://www.fao.org/docrep/008/y5744e/y5744e00.HTM
Hudson, D.A. and Jones, R.G. 2002. Simulation of present day and future climate over
southern Africa using HadAM3H. Hadley Centre Technical Note 38. 37pp.
Hussein, J. 1987. Agro-climatological analysis of growing season in natural regions III,
IV and V of Zimbabwe. In: AGRITEX/GTZ. Proceedings of a Workshop on
Cropping in Semi-arid Areas of Zimbabwe. 24-28 August 1987, Harare.
AGRITEX/GTZ.
Kanemasu, E.T., Stewart, J.I., van Donk, S.J. and Virmani, S.M. 1990. Agroclimatic
approaches for improving agricultural productivity in semi-arid tropics.
Advances in Soil Science, 13: 273-309.
Kinsey, B., Burger, K. and Gunning, J.W. 1998. Coping with drought in Zimbabwe:
Survey evidence on responses of rural households to risk. World Development,
26 (1): 89-110
Love, D., Uhlenbrook, S., Twomlow, S. and van der Zaag, P. 2008. Changing rainfall and
discharge patterns in the northern Limpopo Basin, Zimbabwe. European
Geophysical Union General Assembly, Viena, Austria, April 2008.
http://www.worldwaterweek/stockholmwatersymposium/Abstract_volume_06/
workshop_9.htm.
Mgonja, M.A., Obilana, A.B., Chisi, M., Saadan, H.M., Ipinge, S.A., Mpofu, L., Chintu,
E., Setimela, P., Pali-Shikhulu, J. and Joaquim, E. 2005. Improved sorghum
cultivars released in the SADC region. International Crops Research Institute
for Semi-Arid Tropics (ICRISAT), Bulawayo, Zimbabwe. 60pp.
Monyo, E. 2002. Pearl millet cultivars released in the SADC region. International Crops
Research Institute for Semi-Arid Tropics (ICRISAT), Bulawayo, Zimbabwe.
40pp.
Morse, K. 1996. A Review of Soil and Water management Research in Semi-Arid Areas
of Southern and Eastern Africa. Chatham, UK. Natural Resources Institute. UK.
Mugalavai, E.M., Kipkorir, E.C., Raes, D. and Rao, M.S. 2008. Analysis of rainfall onset,
cessation and length of growing season for western Kenya. Agricultural and
Forest Meteorology, 148: 1123-1135.
26
Mupangwa, W. 2008. Water and nitrogen management for risk mitigation in semi-arid
cropping systems. PhD Thesis (unpublished). University of Free State, South
Africa. 357pp.
Oladipo, E.O. and Kyari, J.D. 1993. Fluctuations in the onset, termination and length of
growing season in northern Nigeria. Theoretical and Applied Climatology, 47:
241-250.
Raes, D., Sithole, A., Makarau, A. and Milford, J. 2004. Evaluation of first planting dates
recommended by criteria currently used in Zimbabwe. Agricultural and Forest
Meteorology, 125: 177-185.
Rockström, J. and Falkenmark, M. 2000. Critical reviews in plant sciences, 19(4):
319-346.
Sivakumar, M.V.K. 1988. Predicting rainy season potential from the onset of rains in
southern Sahelian and Sudanian climatic zones of West Africa. Agricultural and
Forest Meteorology, 42: 295-305.
Stern, R., Dennett, M.D. and Dale, I.C. 1982. Analyzing daily rainfall measurements to
give agronomically useful results. II. A modeling approach. Experimental
Agriculture, 18: 237-253.
Stern, R., Knock, J., Rijks, D. and Dale, I. 2003. INSTAT Climatic Guide. 398pp.
http://www.reading.ac.uk/ssc/software/instat/climatic.pdf.
Sundararaj, D.D. and Thulasidas, G. 1980. The Botany of Field Crops. MacMillan Co.,
India. 508pp.
Tadross, M., Suarez, P., Lotsch, A., Hachigonta, S., Mdoka, M., Unganai, L., Lucio, F.,
Kamdonyo, F. and Muchinda, M. 2007. Changes in growing-season rainfall
characteristics and downscaled scenarios of change over southern Africa:
implications for growing maize. Regional expert meeting report. pp193-204.
www.csag.uct.ac.za/~mtadross.
Unganai, S.L. 1996. Historic and future climatic changes in Zimbabwe. Climate
Research., 6: 137-145.
Usman, M.T. and Reason, C.J.C. 2004. Dry spell frequencies and their variability over
southern Africa. Climate Research, 26: 199-211.
Woltering, L. 2005. Estimating the influence of on-farm conservation practices on the
water balance: Case of the Mzinyathini catchment in Zimbabwe. M.Sc. Thesis.
Delft University of Technology, The Netherlands, 103pp.
27
