I12~>l-?lWW $l',.,,,l, 000
SO261-2194(97)00090-2
PII: ELSEVIER
Dl
Short- and long-term economic implications of controlling crenate broomrape (Orobanche crenata Forsk.) in broad bean (Vicia faba L.) under various management strategies F. Lbpez-Granados*
and L. Garcia-Torres
lnsfitute for Sustainable AgriculturelCSIC, Apdo 4084, 14080 Cdrdoba, Spain
An
economic
controlling under after
model
crenate
different
severity
Crenate
implementing (AB)
respectively. represent
of
Applying
initial
600
herbicide
SOS
herbicide
maximum
was affected
analysis
economic
parameters
(efficacy
sensitivity
coefficients
were between
was
performed
least effect
on AB
treatments
in all subsequent
after
3 and
parasite
herbicide
and applying IS (IS,,)
resulting
at the time
populations
in average for
once
herbicide
Non-application
9 years
of
(Vicia jdw L.)
applying
20.1,
by initial
to determine
of the herbicide,
in all management
at any ISo to minimize
per year
early
of
of herbiannualized
sowing
adequately
of broad
the
of changing
expected
0.5 and I, herbicide strategies.
crenate
regardless effect
dates,
and could only
broomrape
cropping the
values
popula-
frequency. of
the
A
main
yield, fixed and herbicide
costs). In general, which exerted
Generally,
broad bean crops in order
bean
cost being the parameter an IS,,<
1 requires broomrape
to ensure economic
benefits.
0
the
herbicide
1998 Elsevier
Ltd. All rights reserved
Keywords:parasitic weed; planting date; cropping frequency; Introduction
In the Mediterranean region, the Middle East and Eastern Europe the root parasitic weed Orohanche crenatu causes considerable yield losses in legume crops, particularly in broad bean, peas (Pisum sativum L.) and lentils (Lens culinaris Medik.) (Parker, 1994). An integrated control scheme has to include several control measures that are both effective and economically feasible. Reducing the broad bean cropping frequency increases the number of years required to obtain the maximum crenate broomrape IS but not its absolute value (Lbpez-Granados and Garcia-Torres, 1997). Delaying sowing from October to December, although consistently decreasing yield, is a commonly adopted measure for reducing crenate infestation in the Mediterranean area (Raaimakers et al., 1988; Van Hezewijk, 1994). Lowering cropping Fax:
‘)
strategy adopted.
was very low (IS,,
US $ ha I,
against broomrape
net benefits
sensitivity
author
m
implications
bean
low-cost alternative in the long term when IS 2 1 giving an AB of about 653 I after 3 and 9 years for early sowing dates, respectively. Annual application of
was the best strategy
tions and obtain
*Corresponding
[email protected]
in broad
no use of herbicide,
evolution
once did not control
and long-term
an acceptable
and 630 US $ ha
Science
the shortinfestations
broomrape
of the management
infection
and
include
of emerged
population
regardless
when
about
These
number
broomrape
cide was satisfactory benefits
strategies. (IS:
the strategy,
and used to investigate
(Orohanchr crenufu Forsk.)
management
an infection
every year.
is described
broomrape
57
499252,
e-mail:
integrated
control
frequency, delaying sowing and spraying selective herbicide against broomrape, are often included in integrated schemes for broomrape control (Nassib et is available al., 1984). However, no information integrating these all these three approaches on a short and long-term basis. The construction of bio-economic models is a useful technique to assess the economic implications of control strategies. Several authors have made contributions to the evaluation of weed management strategies in order to determine the economic threshold of weed infestation for spraying herbicide or not (Doyle et al., 1986; King et al., 1986; GonzBlezAndlijar and FernBndez-Quintanilla, 1993). Models for the growth of broad bean and its interaction with broomrape and dependence on environmental, and the population dynamics of crenate broomrape in broad bean as affected by planting dates and cropping frequency have been reported (Kropff and Schippers, 1986; L6pez-Granados and Garcia-Torres, 1993a, 1993b, 1997). The objective of this study was
Crop Protection
1998 Volume
17 Number
2
139
Controlling crenate broomrape: F. L6pez-Granados and L. Garcia-Torres to construct an economic model of crenate broomrape growing in broad bean in order to assess the implications of controlling short- and long-term broomrape infestations under several management strategies.
Description
of the model
The presence of crenate broomrape reduces the growth and yield of broad bean crop bean. In selecting the appropriate weed management strategy, the broad bean grower is assumed to be concerned to maximize his annualized benefits (AB, US $ ha- ‘), i.e. the value of the broad bean crop minus the costs of production. The economic model was constructed by competition and the demographic integrating sub-models. Demographic cycle under different sowing dates and cropping frequencies has been closely described by the sigmoidal model: IS, = A/l+ecR(‘P“), where A is the asymptotic value reached at highest weed density, B is the initial IS (time = 0) and C the point of inflection or IS value corresponding to 50% of the maximum IS (LbpezGranados and Garcia-Torres, 1993b, 1997). The annualized benefit (AB) for any control strategy may therefore be expressed as: Ail(t)=
&Y,;,,[
1 - YJlOO( I -f)]
-DF-
DH,
(1)
where BP is the broad bean price (US $ kg-‘), Y,,,Z,X the expected weed-free broad bean crop yield (kg ha-‘), YL the 7o o f yield losses due to broomrape as previously described (YL = 100 IS, 0.124, from MesaGarcia and Garcia-Torres, 1984), f the efficacy of the herbicide treatment, L+ fixed costs and DH herbicide cost. If no herbicide treatment is used the previous equation is simplified as follows: A&t) = BpY,;,x( I -0.124/S,)
- DF.
(2)
Annualized net return (AN) is given by the difference between AB none and AB once or every year. The meaning of all the symbols are given in the following text and in Appendix. The following values have been ascribed to the described parameters according to current prices and costs for a standard farm in Southern Spain in 1997. Thus, BP = US $ 0.27 kg-‘, and Ymax:2800 and 2500 kg ha-’ for early and late planting date, respectively; DE: = US $ 69 ha-’ (pre-tillage sowing, US $ 23 ha- ‘; seed and sowing, US $ 23 ha-‘; and general weed control, $ US 23 ha-‘); DH = US $ 38.5 ha-‘, equivalent to one application of a preemergence and (imazethapyr 75 g ha- ’ from Garcia-Torres 1991) and of a postemergence L6pez-Granados, herbicide treatment (glyphosate 40 g ha-’ from Mesa-Garcia and Garcia-Torres, 1985). The efficacy of the herbicides treatments considered was 90%.
Crenate broomrape
management
strategies
Twelve management strategies were evaluated for the early and late sowing as follows:
140
Crop Protection 1998 Volume 17 Number 2
Early sowing date: 1. Annual broad bean cropping: No herbicide application. la.
Herbicide applied one year after detecting initial infestation. Herbicide applied every year. lc. 2. Triennial cropping: No herbicide application. 2a. Herbicide applied one year after detecting 2b. initial infestation. Herbicide applied every year. 2c. lb.
Late sowing date: 3. Annual broad bean cropping: No herbicide application. 3a. Herbicide applied one year after detecting 3b. initial infestation. Herbicide applied every year. 3c. 4. Triennial cropping: No herbicide application. 4a. Herbicide applied one year after detecting 4b. initial infestation. Herbicide applied every year. 4c. The economic evaluation was made for any of the cropping strategies mentioned above assuming three initial broomrape infestations (IS,,, year 0) of 0.1, 1 and 5 emerged broomrape mP2. However, insofar as the level of broomrape control in the immediate harvest year will also have implications for the parasitic weed levels in subsequent years, it seems reasonable to consider a rather longer period of time, say 3 (short-) and 9 (long-term) subsequent years for consecutive and triennial cropping, respectively. The relevance of the long-run viewpoint is underlined by the observation that many farmers base their decisions about herbicide application on the need to avoid a future build-up of weed populations. Changes in costs and input prices were considered the same for all years. Model sensitivity
to parameter variation
A sensitivity analysis was performed in order to assess the robustness of the AB varying the following
parameters: efficacy of the herbicide (f); broad bean yield level (Y); fixed costs (DF); herbicide cost (D,,). Each parameter value varied by f 15%, except ,f which changed by +6 and - 15%. Sensitivity coefticients (SC) were calculated as the ratio between proportional changes in simulation results (output of the model) and in each parameter: (Aoutputloutput)l(Aparameter/parameter). SC were calculated assuming that initial broomrape population was IS0 = 0.1, 1 and 5. Results
Final broomrape populations (IS), expected average broad bean yield (Y), annualized benefits (AB) and
Controlling crenate broomrape: F. L6pez-Granados and L. Garcia-Torres net returns (AN) are shown in Tuhle I for management strategies described. the twelve Expected AB were 686 and 604 US $ ha- ’ when IS = 0 for early and late sowing date and no herbicide use, respectively (equation 2). annualized
Non-application
of herbicide
IS under any of these strategies rapidly reached very high levels. For every sowing date when ISo = 0.1, the growth rate of the weed population over a 3 and 9-year period was significantly lower, in comparison to higher IS,‘, and final IS was low (between 10 and I8 emerged broomrape m -‘). AB were still high (between 600 and 480 US $ ha ‘), and ensured small broad bean losses (Y > 2 thousand kg ha -‘). population However, when IS,‘> 1, broomrape reached in 3 or 9 years the asymptotic values IS = 62 and 31 for early and late sowing dates, respectively, described by Lopez-Granados and Garcia-Torres (1997) and resulting in a nearly total yield losses (0.3-1.2 thousand kg ha ~ ‘). For example, Y = 0 for early sowing date, every year cropping frequency and IS,, = 5 emerged broomrape mm ‘. Under late sowing and triennial cropping, the maximum IS reached remained low (around 16 emerged broomrape m-‘) and AR were still high (475 and 431 US $ ha-- ‘). For every year cropping the lowest AB was reached, with only 24 US $ ha- ’ after 3 years. One herbicide application Pre or post-emergence herbicides were applied the following year after infections of crenate broomrape were observed. When IS,,< 1, these strategies provided an economically acceptable means of keeping the broomrape population at a low level (IS between 3 and 18) for low initial populations). When IS,, = 0.1, differences for AN were very low under all
of these strategies (between 50 and 73 US $ ha- ‘) because the resulting simulated yield for any cropping frequency (2.7 and 2.3 thousand kg ha- ’ for early and late sowing date, respectively) were very close to the broomrape-free expected yield (2.8 and 2.5 thousand kg ha-’ for early and late sowing date, respectively). Differences between AN were remarkable especially for IS” = 1 for early sowing date and any cropping frequency, giving an AN of 464 and 296 US $ ha-. ‘, respectively. Under these strategies the A13 was considerably higher (namely 565 and 589 US $ ha-‘, respectively) than that obtained with no herbicide application (153 and 293 US $ ha- ‘, respectively) since the broad bean crop achieved its potential yield due to the herbicide use. For IS,, = 5, any sowing date and annual broad bean cropping resulted in a very low yield (0.6 and 1.2 thousand kg ha-’ for early and late sowing date, respectively) and an IS” as high as that obtained with no herbicide treatment (62 and 31 emerged broomrape m m2for early and late sowing date, respectively). Every year herbicide application Under any of these strategies broomrape populations were kept at an unnoticed level (between 0.02 and 0.0003 emerged broomrape m-I), and consequently economic losses were negligible. However, it should be pointed out that for annual broad bean cropping for early sowing date, when initial broomrape infestations were high (I&25), yield was much lower (around 2 thousand kg ha- ‘) than for lower I&. Consequently, maximun AN (536 US $ ha-~ ‘) was reached when IS,’ = 0.1 emerged broomrape rn-~2. Model sensitivity
to parameter
variation
SC are shown in Tut& 2 for IS0 = I emerged broomrape m I, for higher and lower initial populations SC
Table 1. Estimated broomrape populations (IS: number of emerged broomrape m “), average net returns as affected by different initial broomrape infestations and management strategies
broad bean yield, annualized
benefits and
Herbicide application Now
Sowing date Early
Cropping frequency
ISi;
Annual
0.1
I 5 Triennial
Llte
Annual
Triennial
0.1 I 5 0.1
IS”
Every year
Once AB”
IS
IX
2.5 f 0.5
fro0 * 12x
62 62 I6
Oh& 0.03 0+ 0
153_+2hh
4.6 IX
2.7&0,1 2.4 i 0.5
6.53+ 30 565 * 134
IO1 k2flh SOS f 100 2Y3_+ 2YS 2Y3 & 40s 4x0 * 60 23s * I60 24i42 506 *xl 47s +70 431+s7
62 4 I2 2x X.3 I2 31 3 6 I2
0.6 f I, I 2.7 +O.OY 2.3 _+0.3
153f2hh 630 238 sXY+o.l 4.71 fYX SS3i44 513*o4 2S4k IS4 SShi3 541 f Ih SO6 * 5 1
02 02 12
I
31
5 0. I I 5
31 IO I6 I7
2.2 * 0.4 I .2 _+ I.2 0.3 f 0.6 2+0.2 1.2*o.h 0.X f 0.5 2.1 &().I 2*0.3 I .o* 0.2
Y
AB
r
I .o+ 0.1 2.3 kO.2 2.2*0.3 I .2 +o.rl 2.3 + 0.07 2.3 * 0. I 2.2 +0.2
AN’
IS
Y
AR
AN
53 464
0.0003 0.004
2.8 * 0.07 2.x_fo. I
654 + 14 h37_+ I.3
54 536
52 OS 206 I38 73 27X 230 so or, 7s
0.00 0.0003 0.003 0.02 0.00 I2 O.OOh 0.03 0.003 0.003 0.007
2 * O.Oh 2.x & 0. I 2.h + 0.03 2.4 IO.02 2.5 *0.03 2.4 f 0. I 2 50.7 2.5 fO.07 2.5 f 0. I 2.4 + 0.2
4x>+ 12 64X * 4 642ilO ss3+21 570 *Y 5.50+ 2I 4S2k I71 56 I i 20 Shl +x 546 * 2x
35.5 14.7 34’) 200 YO 315 42X 5s x0 115
alnitial crenate broomrape populations bFinal crenate broomrape populations “Average broad bean yield (thousand kg ha ‘) d Average annualized benefits (US $ ha ‘) “Annualized net return (US $ ha ‘) i means the standard deviation of the averaged values ‘Herbicide applied once after IS,, ~0.1 emerged broomrape m *
Crop Protection
1998 Volume
17 Number
2
141
Controlling
crenate broomrape:
F. L6pez-Granados
and L. Garcia-Torres
were similar. In general, the effect of modifying parameter values was low, showing in most cases a SC
Table 2. Sensitivity coefficients [SC = A (V/V)/A (P/P) where V and P are model variable and parameters] for one model output variable, A/3: annualized benefits, to positive and negative changes (%) in herbicide control (f: +6, -15), broad bean yield (V: +15, - 15), fixed costs (/IF +15, - 15) and herbicide cost (& +15, ~ 15). Values of SC were calculated using simulated results from economic model for twelve broad bean management strategies Sowing date
Cropping frequency
Parameter
Herbicide application None
Early
Annual
f Y DI;
1.1 1.1 0.06 0.06
&I Triennial
f
-
Y DF
I .h 0.6 0. I 0.1
DII Late
Annual
f Y DF Dll
Triennial
f
-
Y
1.16 0.9 0.15 0.15 -
DF &I
*Herbicide applied once after &zO.l
142
Crop Protection
1.13 1.3 0.3 0.3 -
Once
Every year
I .os 2.34 0.66 I .49 0.53 0.3 0.4 0.4 I.65 1.1 I.1 1.1 0.1 0.1 0.01 0.02 I .os I .06 1.3 1.1 0.12 0. I2 0.0 I 0.02 0.25 0.55 1.14 1.2 0.13 0.12 0.0 I 0.2
0.06 0.02 1.13 1.13 0.09 0.12 0.03 0.06 0 0.3 I.1 1.1 0. I 0.1 0.05 0.05 0.03 0.05 1.2 1.13 0.13 0.12 0.06 0.05 0.95 0.4 1.2 1.16 0.12 0.13 0.05 0.05
emerged broomrapem *
1998 Volume 17 Number
2
Discussion The economic model described here offers practical guidance for evaluating crenate broomrape management strategies, with the usual limitation of deterministic models. According to our simulated results, when IS,,< 1 early sowing date and annual herbicide application, were the best weed management strategies because crenate broomrape populations were kept under control (IS
1, early sowing dates produced heavy yield losses and crenate broomrape populations promptly reached the asymptotic value reported by L6pezGranados and Garcia-Torres (1997). For the same management strategy, annualized net returns were lower for late sowing date than those obtained in earlier planting dates and this indicates that generally early sowing dates should be recommended.
Controlling crenate broomrape: F. L6pez-Granados and L. Garcia-Torres L.)
Acknowledgements
with
imidazolinones
and
other
herbicides.
Wc~ed Rm.
31.
227-236
This research was partially funded by the Spanish Commision for Science and Technology (CICYT) AGF-950103. We thank Dr GonzBlez-Andtijar for his very helpful comments. Appendix
t
=
time (years)
Variables = initial
Y,,;,,
=
Y YIN
= =
AB
=
4 Dt. Dll
= = =
AN
=
infection severity (emerged broomrapes m-‘) expected weed-free broad bean yield (thousand kg ha ~ ‘) broad bean yield (thousand kg ha ~ ’ ) percentage of yield losses due to broomrape annualized benefits (US $ ha- ‘) broad bean price (US $ kg~ ‘) fixed costs (US $ ha ‘) herbicide costs (US $ ha ‘) annualized net return (US $ ha-- ‘)
B
c
SC
= efficacy of the herbicide
treatment = asymptotic value reached at highest weed density = initial infection severity (time = 0) = point of inflection or infection severity corresponding to 50% of the value maximum infection severity = sensitivity coefficients
References Doyle,
C. J., Cousens, R. and Moss, S. R. (1986) economics of controllingAlopecunrs nlyosuroides wheat. Crop Pmt. 5. 143-150
12. 617-623
Herker, K. N., Blackshaw. R. E. and O’Donovan, J. T. (1905) Field documentation of reduced selection for herbicide resistant weeds with reduced herbicide rates. f’rocecditgs of /tt~ernutiotzul Symposium otl Weed and Crop Rrsisrmce to Herhicider. Cbrdohu. Spaitr, p. 201
Kropff, M. and Schippers, P. (1986) Simulation of the growth of faba beans (Vicia fuhu L.) infested with broomrapes (Orohuncltc c’rmu~u Forsk.) In: Proceedings of u Workshop 017 Biology utld Cotzfrol of Orohuttche. LVIVPO, Wqeningn. The Nc&erlunds. (Ed. by S. J. tcr Borg). pp. 70-79 L6pez-Granados, F. and Garcia-Torres. L. (lYY3a) Seed bank and other demographic parameters of broomrape (Orohunche crmalu Forsk.) populations in fdba bean (Viciu fuhu L.). Wcml Rex 33, 3 I Y-327 Li,pez-Granados, F. and Garcia-Torres, L. ( lY93b) Population dynamics of crenate broomrape (Orohanchc c~rmufrr) in fdba bean (Viciu fuhu). Wbed Sci. 41, 563-567 Wpez-Granados, F. and Garcia-Torres, demography of crcnate broomrape affected by broad bean (Vi& f&u) planting dates. Weed Sci. 45. 37Y-382
L. (1997) Modelling (Orohanche crmutu) cropping frequency
the as and
Mesa-Garcia, J. and Garcia-Torres, L. (1984) A competition index for Orohanche crcnalu Forsk. effects on broad bean (Vi& fuhu L.). Wi,cd Res. 24. 37Y-382
Parameters 5
Spain. Crop Pmt.
Konstantinovic, B. and Milosevic, M. (lYY5) Weed species resistant to herbicide applied in soybean. In: Proceeditlgs qf IntertmGonu1 Symposium on Weed and Crop R~sislut~e fo H~rbicidcs. C‘cirdohu. Spuitl, p. 206
variables
ISO
tion systems in central
King, R. P., Lybecker, D. W.. Scheweizer. E. E. and Zimdahl, R. L. (1986) Bioeconomic modelling to simulate weed control strategies for continuous corn (Zea nzuys). Wrcd Sci. 34, 072-97’)
Definition of symbols Independent
Gonzhlez-Andtijar, J. L. and Fernlindez-Quintanilla, C. (lYY.3) Strategies for the control of Avenu slerilis in winter wheat produc-
A model of the Huds. in winter
Garcia-Torres. L., Cdstej6n-Muimz,
M., Jurado-Expcisito, M. and LBpez-Granados, F. (1996) Modelling the economics of controlling nodding broomrape (Orohunche cernuu Lo&.) in sunflower (Helianthus m7t714145 L.). Weed Sci. 44, S9l-SYS
Garcia-Torres. L. and Lbpez-Grdnados, F. (1991) Control of broomrape (Orohunchc~ crenufa Forsk.) in broad bean (Vi& fuha
Mesa-Garcia, J. and Garcia-Torres, L. (lY85) Orohatzche crenu~a Forsk. control in Viciu fuhu L. with gliphosate as affected by herbicide rates and parasite growth stages. Weed Rm. 25, 12%I34 Nassib. A. M., Hussein, A. H. A. and El Rayes, F. M. (1984) Effect of variety, chemical control, sowing date and tillage on Orohunchc sm. infestation and faba bean vicld. Fuhis Newslerrer 10, II-IS ‘. Parker. C. (1904) The present state of the Orohmche problem. In: Proceedings of rhct Third Internubonul Workshop on Orohunchc utld related Striga research, Royal Tropical InMule. Amsterdam, The N&erluttd.s (Ed. by A. H. Pieterse, J. A. C. Verkleij and S. J. ter Borg), pp. 17-26 Raaimakers, D., Raaijmakers, J., ter Borg, S. J., Nassib, A. M. and Pieterse, A. H. (1988) Effect of sowing date on Orohunche crc’na/u infestation in Viciu fuhu in Egypt. Fuhis Newsl&ter 4. 3%3Y Van Hczewijk, M. J. Relationship Orohutzche (broomrape) development and lentil (Lens culitzuris Medikus) Universiteil Amsterdam pp. h3-7Y. Received 5 May l9Y7 Revised I9 July lYY7 Accepted 8 September
between sowing date and in faba bean (Vicia fuhu L.) in Syria. P/ID Thesis Vrijc
I997
Crop Protection
1998 Volume 17 Number 2
143