Advertisement

Food Security

, Volume 11, Issue 1, pp 69–92 | Cite as

Status quo of chemical weed control in rice in sub-Saharan Africa

  • Jonne RodenburgEmail author
  • Jean-Martial Johnson
  • Ibnou Dieng
  • Kalimuthu Senthilkumar
  • Elke Vandamme
  • Cyriaque Akakpo
  • Moundibaye Dastre Allarangaye
  • Idriss Baggie
  • Samuel Oladele Bakare
  • Ralph Kwame Bam
  • Ibrahim Bassoro
  • Bayuh Belay Abera
  • Madiama Cisse
  • Wilson Dogbe
  • Henri Gbakatchétché
  • Famara Jaiteh
  • Geophrey Jasper Kajiru
  • Alain Kalisa
  • Nianankoro Kamissoko
  • Keita Sékou
  • Ahouanton Kokou
  • Delphine Mapiemfu-Lamare
  • Fanny Mabone Lunze
  • Jerome Mghase
  • Illiassou Mossi Maïga
  • David Nanfumba
  • Abibou Niang
  • Raymond Rabeson
  • Zacharie Segda
  • Fitta Silas Sillo
  • Atsuko Tanaka
  • Kazuki Saito
Open Access
Original Paper

Abstract

If future rice production is to contribute to food security for the increasing population of sub-Saharan Africa (SSA), effective strategies are needed to control weeds, the crop’s fiercest competitors for resources. To gain better insights into farmers’ access to, and use of, herbicides as part of weed control strategies, surveys were conducted in key rice production locations across SSA. Farm surveys were held among 1965 farmers across 20 countries to collect data on rice yields, farmer’s weed management practices, herbicide use, frequencies of interventions and information sources regarding herbicides. Markets were surveyed across 17 countries to collect data on herbicide availability, brand names and local prices (converted to US$ ha−1). Herbicides are used by 34% of the rice farmers in SSA, but adoption ranges from 0 to 72% across countries. Herbicides are more often used by men (40%) than by women (27%) and more often in irrigated (44% of farmers) than in rainfed lowland (36%) or upland rice growing environments (24%). Herbicides are always used supplementary to hand weeding. Following this combination, yield loss reductions in irrigated lowlands and rainfed uplands are estimated to be 0.4 t ha−1 higher than hand weeding alone. In rainfed lowlands no benefits were observed from herbicide use. Sixty-two percent of the herbicides sold at rural agro-chemical supply markets are unauthorized. These markets are dominated by glyphosate and 2,4-D, sold under 55 and 41 different brand names, respectively, and at relatively competitive prices (below average herbicide price of US $17 ha−1). They are also the most popular herbicides among farmers. For advice on herbicide application methods, farmers primarily rely on their peers, and only a few receive advice from extension services (<23%) or inform themselves by reading the product label (<16%). Herbicide application timings are therefore often (38%) sub-optimal. Herbicide technologies can contribute to reduced production losses in rice in SSA. However, through negative effects on crop, environment and human health, incorrect herbicide use may unintentionally counteract efforts to increase food security. Moving away from this status quo will require strict implementation and monitoring of national pesticide regulations and investment in research and development to innovate and diversify the currently followed weed management strategies, agricultural service provision and communications with farmers.

Keywords

Herbicides Glyphosate 2,4-D Farmers Subsistence agriculture Markets 

1 Introduction

Food security in sub-Saharan Africa (SSA) is highly dependent on rice production systems (Seck et al. 2012). Competition from weeds is one of the main biophysical yield constraints in rice production systems in the tropics (Waddington et al. 2010). In SSA, weeds are conservatively estimated to result in annual losses of 2.2 million tons of milled rice (Rodenburg and Johnson 2009). Losses are not only caused by direct resource competition between weeds and the crop, but also because the presence of weeds may attract other biotic yield-reducing factors, such as diseases and grain-feeding birds (Heinrichs et al. 1997; Demont and Rodenburg 2016). Furthermore, while weed-inflicted yield losses may be diminished through weed control, these efforts depend on inputs such as labor (Ogwuike et al. 2014) which in turn imply additional indirect economic losses.

Previous studies have estimated yield loss reductions of at least 1 t ha−1 following improved weed management (Haefele et al. 2000; Becker et al. 2003; Nhamo et al. 2014). However the efficacy of actual weed management — in terms of yield loss prevention — in SSA is among the lowest in the world (Oerke and Dehne 2004). At farm level, weed-inflicted yield losses, despite control efforts by the farmer, were estimated to be still 15% in irrigated lowlands, 16% in rainfed uplands and 23% in rainfed lowlands (Becker and Johnson 2001a, b; Becker et al. 2003).

Weeds are mainly controlled manually, mechanically or chemically. The first option, hand weeding, is labor intensive; in upland rice systems hand weeding was estimated to take 173 to 376 person-hours per hectare, depending on the number of weeding interventions (Ogwuike et al. 2014). Mechanical tools for weeding, either person-driven, animal-driven or engine-driven, are scarce in rice systems in Africa (Rodenburg and Johnson 2009; Gongotchame et al. 2014), despite a latent interest from farmers (Johnson et al. 2018). Herbicide application, when applied well, is usually the most effective and least labor-intensive weed control method with the highest yield return (Rodenburg et al. 2015). This technology, however, relies heavily on the availability of well-functioning agro-chemical supply markets as well as on sufficient financial means and know-how on application techniques at the level of the farmer or service provider. These preconditions are often not met in smallholder rice systems in rural areas in SSA (Balasubramanian et al. 2007). Therefore, adoption rates of herbicides in SSA are estimated to be low (Gianessi 2013), and application characteristics i.e. herbicide choice, rate and timing, are assumed to deviate frequently from the recommendations (Rodenburg and Johnson 2009) with potential negative consequences for the environment, human health, crop performance (Zimdahl 2007) and hence food security.

Data on weed management practices, herbicide availability, prices and use in SSA are scarce (Rodenburg and Johnson 2009). This information gap in turn complicates the identification of entry points for innovations in weed control in African smallholder rice production systems. The objectives of this study were therefore, through a survey, to (1) assess the current importance of herbicides in weed management strategies of smallholder rice farmers in SSA, (2) find out whether this technology could contribute to productivity enhancement and therefore food security, (3) assess the availability and prices of rice herbicides on rural markets in SSA, (4) learn what types of herbicides are used by farmers, (5) find out what the sources of information are that farmers use concerning herbicide application, and (6) discover how all this is reflected in the actual use of these products in farmers’ fields.

2 Materials and methods

2.1 Site and farmer selection

Farmer-surveys were conducted in 222 randomly selected villages in 36 sites divided over 20 countries in SSA (5 in East Africa: Ethiopia, Madagascar, Rwanda, Tanzania and Uganda, 12 in West Africa: Benin, Burkina Faso, Côte d’Ivoire, Ghana, Guinea, Mali, Niger, Nigeria, Senegal, Sierra Leone, The Gambia and Togo, and 3 in Central Africa: Cameroun, DR Congo and Chad). In each country, sites were selected by the National Agricultural Research Institute (NARI) and its partners. Selected sites were considered priority intervention sites for national rice research and development (see: Niang et al. 2017; Tanaka et al. 2017). The sites covered five tropical agro-ecological zones: the arid zone (AR), the semi-arid zone (SA), the sub-humid zone (SH), the humid zone (HU) and the highland sub-humid zone (HL) (3.1.1, 3.1.2, 3.1.3, 3.1.4 and 3.2.3, respectively, of the classification by HarvestChoice 2010) (Fig. 1, Table 1). Most sites (27) are characterized by one rice-growing environment (either irrigated lowlands, rainfed lowlands or rainfed uplands). The remaining sites encompassed irrigated lowlands and rainfed lowlands (Lagdo in Cameroun, Savelugu in Ghana, Gaya in Niger, Kahama in Tanzania), rainfed lowlands and rainfed uplands (Glazoue in Benin, Sikasso in Mali, Kilombero in Tanzania, Namulonge in Uganda), or even all three environments (Navrongo in Ghana). Site categorization in either irrigated lowlands, rainfed lowlands or rainfed uplands was done by national experts of the NARIs. Location-specific names of production environments ‘inland valley swamps’ and ‘riverine’ in Sierra Leone were classified as rainfed lowland.
Fig. 1

Locations of the 36 study sites, in 20 countries, where the farmer-surveys were conducted (2012-2014), overlapped by the agro-ecological zones as defined and mapped by HarvestChoice (2010). The market study was conducted in site numbers 1-3, 5-15, 19, 21, 23-27 and 31-35

Table 1

Information on the survey conducted: agroecological zones (AEZ), countries, site names, villages (V), altitude range, and the number of surveyed farmers distributed over gender, rice growing environments and rice establishment methods.

AEZ

Country

Site

MS

V

Altitude (m)

Gender

Environment

Planting

Min.

Max.

F

M

IL

RL

RU

DS

TP

SA

Benin

Malanville

*

3

164

231

17

58

75

0

0

0

75

SH

 

Glazoue

*

7

155

249

34

29

0

34

29

63

0

SH

B. Faso

Cascades

*

5

272

361

23

21

0

44

0

33

11

SH

 

Hauts-bassins

 

2

0

21

0

0

21

21

0

HL

Cameroon

Ndop

*

4

1127

1202

30

20

0

50

0

1

49

SH

 

Lagdo

*

7

145

258

20

44

21

43

0

40

24

SA

Chad

Tandjilé-Est

*

5

14

34

0

48

0

48

0

HU

Côte d’Ivoire

Gagnoa

*

6

169

270

3

52

55

0

0

18

37

HU

 

Man

*

5

324

460

23

50

0

0

73

73

0

SH

DR Congo

Bandundua

*

5

279

450

2

40

0

0

42

42

0

HL

Ethiopia

Fogera

*

5

11

29

0

40

0

39

1

SA

Gambia

Central River

*

6

0

125

31

39

70

0

0

12

58

SA

 

West Coast

*

5

12

63

33

37

0

0

70

55

15

HU

Ghana

Kumasi

*

4

225

247

7

27

0

34

0

22

12

SH

 

Afife

 

1

22

57

7

42

49

0

0

27

22

SH

 

Navrongo

*

8

95

336

28

58

32

42

12

65

21

SH

 

Savelugu

 

9

64

170

6

88

6

88

0

88

6

SH

Guinea

Haute Guinée

 

5

370

500

2

66

0

0

68

68

0

HL

Madagascar

Ambohibary

*

6

1545

1693

16

45

0

61

0

7

54

HL

 

Ankazomiriotra

 

4

1045

1180

3

40

0

0

43

36

7

SA

Mali

Kouroumari

 

5

299

319

8

45

53

0

0

6

47

SA

 

Sikasso

*

6

340

385

61

38

0

89

10

87

12

SA

Niger

Gaya

*

5

170

191

3

46

11

38

0

3

46

SA

 

Tillabery

*

5

210

269

1

64

65

0

0

9

56

SH

Nigeria

Nasarawa

*

5

116

209

9

42

0

51

0

42

9

HL

Rwanda

Rugeramigozib

*

9

1760

1797

35

15

50

0

0

0

50

HL

 

Rwasave

*

1

1595

1804

21

29

50

0

0

0

50

AR

Senegal

Dagana

 

5

0

23

4

37

41

0

0

30

11

SH

S. Leone

Bo & Kenema

 

36

2

219

13

46

0

59

0

16

43

SH

 

Tormabum

 

6

5

14

10

40

0

50

0

41

9

SH

Tanzania

Kahama

*

5

1146

1208

15

41

8

48

0

13

43

SH

 

Kilombero

*

3

211

266

9

12

0

20

1

18

3

SH

Togo

Rég. Plateaux

*

3

257

328

21

27

0

48

0

25

23

SH

 

Rég. Maritime

*

3

39

75

10

29

39

0

0

7

32

HU

Uganda

Dohoc

*

7

1054

1102

6

28

34

0

0

0

34

HU

 

Namulonge

 

16

1028

1132

7

43

0

38

12

50

0

 

20

36

26

222

0

1804

543

1422

659

925

381

1105

860

AEZ refers to agro-ecological zones: AR arid, HL highland, HU humid, SA semi-arid, SH semi-humid; Environment refers to rice growing environment: IL irrigated lowland, RL rainfed lowland, RU rainfed upland; Planting refers to crop establishment method: DS direct seeded, TP transplanted; MS refers to market study, with * indicating sites where the market study was conducted; V refers to village, indicating the number of villages per site; Gender differentiates men (M) from women (F) farmers

aKinshasa, Bas-Congo

bRugeramigozi is also known as Gikonko II and Rwasave is known as Gikonko I

cEastern Uganda

2.2 Farm surveys

In each site the target community and farmers were selected by a team of socio-economic researchers of the local NARI and AfricaRice, following standardized protocols (see: Niang et al. 2017; Tanaka et al. 2017), whereby the leading criterion was for a farmer to manage at least one rice production field with a minimum size of 200 m2. Attempts were made to come to a farmer selection that was considered representative of a specific site. The sample size (number of farmers per community × number of communities surveyed) depended on the number of technicians available, the technicians’ experience in field surveying, the budget, and the size of the site.

A rapid rural appraisal (RRA) was held on weed management practices during the wet seasons of 2012, 2013 or 2014 among 1965 individual rice farmers, using a structured questionnaire. Basic information was gathered in each site: (1) village names, (2) minimum and maximum altitude, (3) gender of participating farmers, (4) rice-growing environment the participating farmer operates in, and (5) the crop establishment method a farmer follows i.e. transplanting or direct sowing (i.e. dry-seeding in uplands, wet-seeding in lowlands). Leading follow-up questions of the farmer-survey were: (1) what weed management strategies do you apply (hand weeding, mechanical weeding, herbicide application), (2) how often do you conduct a weed management intervention during a season, (3) if you use herbicides, what kind of herbicides do you use, (4) do you apply herbicides yourself, and (5) if you apply them yourself, how do you obtain information on application methods (product label, agricultural extension, neighbor/colleague, other). The questions on weeding methods, herbicide choice and application timing (pre- or post- weed emergence), were asked for each weeding intervention stage within the season, including “Weeding after land preparation, but before crop establishment” (W1), “1st weeding after crop establishment” (W2), “2nd weeding after crop establishment” (W3), “3rd weeding after crop establishment” (W4) and “4th weeding after crop establishment” (W5). This structured way of questioning reduced the likelihood of misconceptions between the enumerator and the farmer.

In the survey, the term hand weeding referred to the practice of uprooting weeds by hand, often combined with the use of a short-handled hoe, and removing them from the field by hand. Mechanical weeding referred to weeding operations that only made use of mechanical implements, either hand- animal- or fuel-driven. These implements included machetes, push or rotary weeders, sine hoes, harrows, spades, oxen ploughs, power tillers and tractor-mounted harrows.

2.3 Rice yield assessments

At each farm where a farmer survey was conducted, the rice (paddy) yield was assessed from three 12 m2 harvesting areas (3 m × 4 m) that were randomly assigned to the field. Panicles were cut and threshed, and the collected grains were then winnowed and weighed. Grain moisture content was measured at the time of weighing using digital grain moisture meter (SATAKE Eng. Co., Tokyo; Model SS-7) to correct the grain weight to a standard moisture content of 14%. Grain weights were then extrapolated to tons of paddy ha−1.

2.4 Market surveys

Between 2014 and 2015, an additional survey was conducted at markets among a sub-set of 26 sites (see Table 1), out of the 36 sites where farmer-surveys were held. This survey covered 17 countries. Of the 20 previously mentioned countries only Guinea, Senegal and Sierra Leone were not included in this survey. In each location at least three agro-chemical shops were visited and in each of these shops all available herbicide brands were listed. For each brand, country of origin, company information and recommendations concerning application provided on the label were noted down. In addition, for each herbicide brand the local price per bottle (local currencies) as well as the bottle volume was noted. Herbicide prices were converted from the local currency to US dollars ($) and from the bottle price to a price per ha, following recommended application rates.

Public consultations of available information sources were made to check for nine countries whether herbicide products were authorized before the date of the market study. For Benin, Burkina Faso, Chad, Côte d’Ivoire, Mali, Niger, The Gambia and Togo the list of herbicides authorized by the Comité Sahélien des Pesticides (CSP 2013) was used, while for Tanzania the list of registered pesticide products in Tanzania, of the Tropical Pesticide Research Institute (TPRI 2011), was used.

2.5 Data analyses

Descriptive statistics were generated for weed management practices, weeding timing and frequency, herbicide prices, herbicide types and sources of information for herbicide use. Where relevant, data were disaggregated by gender and/or rice growing environment. Pearson Chi-square (χ2) tests of independence were performed to determine whether there were significant relationships (P < 0.05) among the data (i.e. number of weeding interventions, weed management practices and herbicide information sources) and gender (men, women) and/or environment (irrigated lowlands, rainfed lowlands, rainfed uplands). Two linear regression analyses were conducted to quantify variation in rice yields due to (1) the weeding frequency and (2) the weeding method. Both these analyses were done for each rice growing environment separately, but across agro-ecological zones. Weeding frequencies ranged from no-weeding (W0) to four or more weeding interventions (W4+), whereby W0 was used as the reference. The W4+ category comprised farmers following four, five and more than five weeding interventions because the sample sizes of these categories on their own were relatively small. Based on the available data, the weeding method followed by farmers comprised four categories: hand weeding only (HW), hand weeding and herbicide application (HW + H), hand weeding and mechanical weeding (HW + M), and hand weeding and herbicide application and mechanical weeding (HW + H + M). The first category (HW) was used as the reference situation. Coefficients of these two regressions, that represent the variation in the rice yields when switching from the reference category to any of the other categories, were estimated with their associated standard errors and P value. All data analyses were done using R software, Version 3.4.1 (R-Core-Team 2017).

3 Results

3.1 Farm characteristics

The altitude of rice production areas ranged from sea level to 1804 m above sea level (Table 1). The farmer selection was composed of 38% women and 62% men. The majority of farmers were producing rice in the rainfed lowlands (47%) followed by irrigated lowlands (34%) and rainfed uplands (19%). Slightly more than half of the farmers (56%) established their rice crop by direct sowing, whereas the others (44%) transplanted rice seedlings from nurseries. The importance of transplanting depends on rice growing environment. In irrigated lowlands 76% of the rice crops were established by transplanting, while in rainfed lowlands this was only 37%. In the uplands, all rice was established by direct sowing.

3.2 Weed management practices

The majority of rice farmers intervened only once (34%) or twice (39%) during a season to control weeds (Fig. 2). Only 5% of the farmers did not intervene at all, while 22% intervened three times or more. There was a significant difference in weeding intervention frequency between men and women (χ2: 16.21; P = 0.0063). Compared with men, a higher share of women intervened twice, and a lower share of women intervened three times or more (Fig. 2a). There was also a significant difference in weeding intervention frequency between rice growing environments (χ2: 95.99; P < 0.0001), in particular between the rainfed (both upland and lowland) and the irrigated environments. Compared with farmers in irrigated lowlands, a higher share of farmers in rainfed rice fields intervened only once or not at all, and a lower share intervened two or three times (Fig. 2b).
Fig. 2

Distribution of number of weeding interventions per farmer category, a Gender: men (n = 1422) or women (n = 543), χ2 = 16.21; P = 0.0063, b Environment: irrigated lowland (n = 659), rainfed lowland (n = 920) or rainfed upland (n = 380), χ2 = 95.99; P < 0.0001

Weeding resulted in yield loss savings in all rice growing environments but the extent of these savings depended on the number of weeding interventions. In none of the environments was the yield loss reduction obtained after just one weeding significant (Table 2). In irrigated lowlands weeding significantly reduced yield losses after two or more weeding interventions. Compared with farmers who did not weed at all, farmers who weeded twice or three times gained around 1 t ha−1. Close to 2 t ha−1 of yield loss reductions were estimated to be obtained in this environment by farmers intervening four or more times. In rainfed lowlands, weeding effects were smaller and less consistent. Farmers who weeded twice obtained an estimated 0.6 t ha−1 of yield loss reductions while farmers weeding three times did not see a significant positive effect. Farmers weeding four times or more obtained an estimated 1.5 t ha−1 yield savings. In rainfed upland, farmers weeding twice obtained an estimated yield loss reduction of 0.4 t ha−1, and this improved to just below 1 t ha−1 after weeding three times and 1.2 t ha−1 after weeding four times or more (Table 2).
Table 2

Regression analyses output, quantifying the variation in rice yields due to the (1) Weeding Frequency and (2) the Weeding Method

 

Environ.

Level

n

Estimate

SE

P

Weeding frequency

ILa

(Intercept: W0)

14

2.999

0.428

<0.0001

W1

160

0.580

0.448

0.1954

W2

327

0.934

0.437

0.0329

W3

113

1.045

0.454

0.0218

W4+

45

1.970

0.491

<0.0001

RLb

(Intercept: W0)

65

2.131

0.216

<0.0001

W1

345

0.196

0.235

0.404

W2

293

0.568

0.240

0.018

W3

141

0.301

0.259

0.246

W4+

76

1.512

0.290

<0.0001

RUc

(Intercept: W0)

24

1.159

0.190

<0.0001

W1

161

0.211

0.204

0.302

W2

136

0.403

0.206

0.052

W3

45

0.963

0.235

<0.0001

W4+

14

1.249

0.321

0.0001

Weeding method

ILd

(Intercept: HW)

239

3.725

0.107

<0.0001

HW + H

247

0.377

0.148

0.0109

HW + M

123

0.039

0.182

0.829

HW + H + M

36

1.070

0.288

0.0002

RLe

(Intercept: HW)

439

2.567

0.083

<0.0001

HW + H

290

−0.232

0.132

0.0783

HW + M

115

0.684

0.176

0.0001

HW + H + M

11

0.059

0.507

0.908

RUf

(Intercept: HW)

191

1.495

0.073

<0.0001

HW + H

75

0.397

0.136

0.0037

HW + M

84

0.044

0.130

0.7338

HW + H + M

6

−0.502

0.407

0.2187

Both analyses are broken down in rice growing environments: irrigated lowland (IL), rainfed lowland (RL) and rainfed upland (RU). Weeding Frequency has four levels, ranging from no-weeding (W0) —taken as reference— to four or more weeding interventions (W4+); Weeding Method has four levels: Hand Weeding (HW) —taken as reference—, Hand Weeding + Herbicides (HW + H), Hand Weeding + Mechanical Weeding (HW + M) and Hand Weeding + Herbicides + Mechanical Weeding (HW + H + M). Output shows the regression coefficients (Estimate, i.e. estimated yield or yield changes in t ha−1), standard errors (SE) and P-values showing significance of contributions

aResidual standard error (RSE): 1.601; Degrees of freedom (DF): 637 degrees of freedom (DF); Observations deleted due to missing data (MD): 17

bRSE: 1.648; DF: 838; MD: 77

cRSE: 0.931; DF: 363; MD: 12

dRSE: 1.605; DF: 624; MD: 17

eRSE: 1.657; DF: 781; MD: 70

fRSE: 0.982; DF: 340; MD: 12

The most commonly used weed management practice by rice farmers was hand weeding (mean: 93%; weighted mean: 95%), followed, by a wide margin, by herbicide application (mean: 31%; weighted mean: 34%) and mechanical weeding (mean: 16%; weighted mean: 21%) using sine hoes, machetes, push or rotary weeders (Table 3). Herbicide use or mechanical weeding was always combined with hand weeding but only 3% of the farmers combined herbicides with mechanical weeding (not shown). Men (40%) used herbicides significantly (P < 0.001) more often than women (27%), while no gender differences were observed between the use of hand weeding or mechanical weeding (Fig. 3a). Both hand weeding and herbicides were significantly (P < 0.001) more frequently applied in irrigated lowlands, compared with upland or lowland rainfed rice environments (Fig. 3b). Herbicides were applied by 44% of the farmers in irrigated lowlands, compared with 36% in rainfed lowlands and only 24% in rainfed uplands. In rainfed lowlands a significantly (P < 0.001) lower share of the farmers used mechanical tools for weed control compared with farmers in irrigated lowland and rainfed uplands. There was also wide variation in the type of weed control interventions across countries (Table 3), in particular in herbicide use (ranging from 0 to 72%) and mechanical weeding (0 to 84%). In 12 countries, herbicide use ranged between 32 and 72% (with a mean of 51%) while in the remaining 8 countries (Ethiopia, Chad, Madagascar, The Gambia, Tanzania, DR Congo, Rwanda and Sierra Leone), herbicide use was only 3% or less. In The Gambia, and Madagascar this near absence of herbicide use was compensated by a high rate of mechanical control (>80%). In 11 of the 20 countries, mechanical weed control was not practised at all. This included five of the countries where farmers also hardly used herbicides (i.e. Ethiopia, Chad, DR Congo, Rwanda and Sierra Leone).
Table 3

Percentage of farmers using different weed control interventions in the surveyed countries

  

Type of weed control intervention

Country

n

Hand

Mechanical

Chemical

Benin

138

100

46

59

Burkina Faso

65

97

0

72

Cameroon

114

89

2

43

Côte d’Ivoire

128

99

1

41

Chad

48

100

0

2

DR Congo

42

60

0

0

Ethiopia

40

100

0

3

Gambia

140

100

84

1

Ghana

263

95

36

55

Guinea

68

90

0

49

Madagascar

104

99

82

2

Mali

152

93

0

43

Niger

114

100

15

58

Nigeria

51

84

0

37

Rwanda

100

100

0

0

Senegal

41

88

0

68

Sierra Leone

109

84

0

0

Tanzania

77

92

28

1

Togo

87

100

0

57

Uganda

84

96

17

32

Mean

 

93

16

31

Weighted mean

 

95

21

34

Min

 

60

0

0

Max

 

100

84

72

Fig. 3

Percentages of farmers managing weeds by hand, herbicides or mechanical implements in each of the farmer category, a Gender: women (n = 512) or men (n = 1322), and b Environment: irrigated lowland (n = 645), rainfed lowland (n = 841) or rainfed upland (n = 342); Indications of significant effects following Chi-square (χ2) tests refer to comparisons within categories of weed management practices, whereby ‘ns’ means not significant and *** indicates significance at P < 0.001; Gender: χ2Hand weeding = 1.020; χ2Chemical control = 26.124; χ2Mechanical weeding = 2.341; Environment: χ2Hand weeding = 1.020; χ2Chemical control = 39.603; χ2Mechanical weeding = 2.341

Farmers in irrigated lowlands obtained significantly higher yield loss reductions from their weeding efforts when they supplemented hand weeding with herbicide applications (0.4 t ha−1) or a combination of herbicide applications and mechanical weeding (1.1 t ha−1; Table 2). In rainfed lowlands farmers supplementing hand weeding with mechanical weeding obtained significant (0.7 t ha−1) higher yields than farmers pursuing hand weeding only. Compared with the control group (hand weeding only), farmers who supplemented hand weeding with herbicides obtained an estimated 0.2 t ha−1 lower yields, while farmers combining all three methods obtained no significant yield advantage. In rainfed uplands supplementing hand weeding by herbicide application resulted in a significant yield advantage (0.4 t ha−1), while combining hand weeding with mechanical or combinations of chemical and mechanical weed control did not result in significantly different yields compared with hand weeding only.

3.3 Chemical weed control

Herbicides were mainly applied during the first weeding intervention, before crop establishment, only (W1: 61% of cases), during the first and second weeding intervention (W1,2: 26%) or during the second intervention only (W2: 7%) but there were differences among rice growing environments (Fig. 4). Compared with rainfed environments, in irrigated lowlands a relatively larger share of farmers applied herbicides at more or later stages than before crop establishment only (W1). Compared to other environments, in rainfed lowlands the share of farmers applying at W1 only was the highest, while in the upland relatively more farmers applied herbicides during W1,2, and the first four weeding interventions (W1,2,3,4; Fig. 4).
Fig. 4

Frequency and timing of herbicide application for all farmers applying herbicides (large chart left; n = 719) and per rice growing environment (smaller charts right): Irrigated lowland (n = 333), rainfed lowland (n = 301) and rainfed upland (n = 81). Categories indicate the weed intervention periods, with W1 = After land preparation but before crop establishment, W2 = 1st intervention after crop establishment, W3 = 2nd intervention after crop establishment, W4 = 3rd intervention after crop establishment

The 677 herbicide users (34%) used a total of 18 known herbicide formulations (Table 4). The most frequently used formulations were glyphosate (with 176 farmers at W1, 45 farmers at W2 and 4 farmers at W3), 2,4-D (118 farmers at W1, 47 at W2 and 4 at W3) and bensulfuron (112 farmers at W1, 38 at W2 and 7 at W3). In 39 cases (20 at W1, 2 at W2 and 17 at W3) farmers were not able to tell the name of the herbicide formulation used (Table 4, ‘unknown’).
Table 4

Herbicide application according to farmers; Number of farmers applying herbicides prior to crop establishment (W1), first application after crop establishment (W2) and second application after crop establishment (W3); Farmers were asked whether the herbicide was applied pre-emergence (PRE) or post-emergence (POST) of the weeds

Formulations

Type

Number of farmers reporting

Applied before crop establishment (W1)

First intervention after crop establishment (W2)

Second intervention after crop establishment (W3)

Potential misuse frequencya

Potential misuse percentage

PRE

POST

PRE

POST

POST

Glyphosate

PPb

225

139

37

28

17

4

188

84

2,4-D

POST

169

15

103

7

40

4

22

13

Bensulfuron

POST

157

46

66

11

27

7

57

36

Propanil +2,4-D

POST

60

1

40

 

16

3

1

2

Butachlor

PRE

43

18

4

17

3

1

8

19

Propanil

POST

39

 

20

1

15

3

1

3

Propanil + Triclopyr

POST

26

2

19

2

1

2

4

15

Pendimethalin

PRE

23

16

3

3

1

 

4

17

Paraquat

PP

15

8

7

     

Metolachlor + Terbutryn

PRE

10

3

1

 

6

 

7

70

Propanil + Butachlor

PREc

8

 

8

   

8

100

Pretilachlor + Pyribenzoxim

POST

4

   

3

1

0

0

Oxadiazon

PRE

2

1

1

   

1

50

Haloxyfop

PRE

1

 

1

   

1

100

Paraquat + Pendimethalin

PRE

1

 

1

   

1

100

Trifluralin

PRE

1

1

    

0

0

Propanil + Bentazon

POST

1

 

1

   

0

0

Glyphosate + Oxyfluorfen

PRE

1

   

1

 

1

100

Unknown

 

7

13

1

1

17

(39)d

 

Total number

 

786

257

325

70

131

42

312

38

aPotential misuse frequency (%) is the number of cases where the herbicide application timing, indicated by the farmer, does not match with the recommended herbicide application timing, indicated on the product label, divided by the total number of herbicide applications and multiplied by 100

bPP = Pre-planting, the herbicide should be applied pre planting but post weed emergence, the assumed correct use is therefore only under W1 as POST

cCan be applied early post-emergence

dFamers not knowing the product, indicate a lack of awareness and a high risk of misconception/misuse, or concern cases where herbicide application was outsourced to service providers (like in Rwanda). These cases (‘unknown’) were not included in the calculation of the potential misuse frequency

Farmers more often applied herbicides post weed emergence (W1: 56%; W2: 65%) than pre-emergence (Table 4). Based on the farmers’ responses, we estimated the proportion of potentially wrong application timings at 38% (Table 4). This ‘potential misuse frequency’ is the number of cases where the herbicide application timing, indicated by the farmer, did not match with the recommended timing of herbicide, indicated on the product label, divided by the total number of herbicide applications. Of the top-ten most frequently used herbicides, the highest percentage of potential misuse was observed with glyphosate (84% of cases where it was applied), followed by metolachlore + terbutryn (70%) and paraquat (53%).

For farmers, the most consulted sources of information on herbicide application were neighboring farmers (Table 5). Some 63% of the women and 66% of the men turned to their neighbor for advice during the first weeding intervention (W1) and 56% of the women and 63% of the men solicited neighbor advice during the second intervention (W2). At any intervention time, no more than 23% of the farmers sought advice from extension services and no more than 16% read the herbicide product labels before applying herbicides (Table 5). When broken down to countries (only those with at least 30% of farmers using herbicides), the data show that extension services were important information sources in Togo, Mali and Senegal, while in Uganda farmers mostly relied on the product label and farmers in Burkina Faso and Nigeria more often consulted their neighbors (Fig. 5).
Table 5

Sources of information for the farmers using herbicides during the first weed management intervention time (W1) and the second weed management intervention time (W2)

  

Product label

Agricultural Extension

Neighbor farmer

Other sources

Women

Men

Women

Men

Women

Men

Women

Men

W1

Consulteda

11

12

23

18

63

66

2

4

Not-consulted

89

88

77

82

37

34

98

96

χ 2 b

0.091 (ns)

1.669 (ns)

0.377 (ns)

0.385 (ns)

W2

Consulted

16

14

21

16

56

63

2

5

Not-consulted

84

86

79

84

44

37

98

95

χ 2

0.009 (ns)

0.236 (ns)

0.581 (ns)

0.183 (ns)

aResponses are independent, meaning that multiples sources may have been consulted by the same farmer

bPearson’s Chi-square (χ2) test statistic indicating a gender effect on the way farmers inform themselves on the application practices of herbicides; ns = not significant

Fig. 5

Information sources of herbicide-using farmers (HU) (Extension services, Product label, Neighboring farmers, None) for the application of herbicides, broken down per country for the countries with more than 30% of herbicide users (N = 713): Togo (HU = 57%), Uganda (HU = 32%), Cameroun (HU = 43%), Mali (HU = 43%), Senegal (HU = 68%), Côte d’Ivoire (HU = 41%), Niger (HU = 58%), Guinea (HU = 49%), Benin (HU = 59%), Ghana (HU = 55%), Burkina Faso (HU = 72%), and Nigeria (HU = 37%). Values between parentheses behind country names represent the number of herbicide users (n). The category ‘other sources’ was too insignificant to be shown here

3.4 Herbicides on the market

During the market-survey, in 26 locations, a total of 235 herbicide brands were encountered and examined. These brands encompassed 17 different herbicide formulations, 11 single-formulations and 6 combined-formulations (Table 6). By far the most common formulation found on the market was glyphosate. The formulation was found in 42% of the herbicides. Second, 2,4-D amine, was found in 33% of the herbicides (stand-alone: 27%, combined: 6%) and the third formulation, propanil, was found in 16% of the herbicides (stand-alone: 2%, combined: 14%). Herbicide formulations were available under a wide diversity (150) of brands, in particular glyphosate, with 55 different brands, and 2,4-D, with 41 brands (Table 6). In some cases different herbicide formulations were sold under the same or very similar brand names (although not at the same location); e.g. Weed Kill for 2,4-D in Cameroon and for glyphosate in Uganda and Weed Killer for 2,4-D in Ethiopia.
Table 6

Formulations and brand names of the herbicides available in agrochemical supply shops in rice growing areas of SSA

 

Herbicide brands

Formulations

Number of observations

Frequency

Number

Examples

Glyphosate

98

41.7

55

Adwumaye; Agasate; Agriherb; Destroyer; Detru-Herb; Frankosate; Gly Star; Glycel; Glycot; Glyfort; Glyphader; Herbo Total; Heros; Kalach; Lamachette; Nwura Wura; Puissance; Roundup; Sunphosate; Tackle; Touch Down; Uproot; Weed Kill; Weedall

2,4-D

63

26.8

41

Agriselect; Amino Force; Ascomine; Bextra; Cotomine; Dekade Plus; Devaweed; Herbafor; Herbazol; Herbextra; Herbus Plus; Hond; Stopstar; Sun; Ultra 2;4-D; Weed Kill; Weed Killer

Propanil + Triclopyr

18

7.7

10

Calriz; Garil; Maloflora; Phytoriz; Pyranyl; Rigold; Rivitex; Sakaril; Tripro; Tropiryle

Propanil +2,4-D

13

5.5

8

Baccara; Orizo Plus; Pronil Plus; Propa Gold; Propa Plus; Propacal Plus; Propocalpus; Vespanil Plus

Paraquat

7

3.0

6

Gramoquat Super; Gramoxone; Kabquet; Para Q; ParaCot; Weed Crusher

Pendimethalin

7

3.0

5

Activus; Alligator; Kayanga; Pendimenthalin; Stomp

Butachlor

5

2.1

5

Buta Force; Butachruseh; Butaplus; Surplus; Ultrachlore

Propanil

4

1.7

4

Propanil; Propanil Plus; Propercare; Yuperstar

Oxadiazon

4

1.7

4

Topstar; Callistar; Oxariz; Ronstar

Bensulfuron

4

1.7

3

Condax; Dadyax; Samory

Bispyribac

3

1.3

2

Bounty; Nominee Gold

Bensulfuron + Pretilachlor

2

0.9

1

Londax

Haloxyfop-R

2

0.9

1

Gallanfort

Propanil + Thiobencarb

2

0.9

2

Herbivore; Rical

Pretilachlor + Pyribenzoxim

1

0.4

1

Solito

Penoxsulam

1

0.4

1

Rainbow

Glyphosate + Oxyfluorfen

1

0.4

1

Zoomer

Total

235

100

150

 
Apart from being the most widely available formulations, 2,4-D ($10 ha−1; n = 60) and glyphosate ($15 ha−1; n = 97) were also among the cheapest herbicides on the market (Fig. 6a). Mean prices of the third and fourth most widely available herbicides; i.e. combinations of propanil + triclopyr ($48 ha−1; n = 18) and propanil +2,4-D ($25 ha−1; n = 12), were both well above the overall mean herbicide price of US $17.4 ha−1. Large price differences among countries were observed (Fig. 6b). Mean herbicide prices in Mali ($26 ha−1; n = 7), The Gambia ($25 ha−1; n = 8) and Benin ($24 ha−1; n = 22), were above average, while in Ethiopia ($4 ha−1; n = 4), Tanzania ($9 ha−1; n = 9), Uganda ($11 ha−1; n = 5) and Ghana ($11 ha−1; n = 36) they were well below average.
Fig. 6

Herbicide prices per ha per product (a) and per country (b) with the number of observations per product or country (n) between parentheses. Black dots indicate means, error bars indicate standard error of means, hyphens indicate minimum and maximum values and the dashed line indicates the weighted mean (US $17.4 ha−1). Absence of some herbicide products is caused by missing or incomplete data for conversion in price per ha

Of the available herbicide brands on the market, 62% appeared not to be authorized by a recognized pesticide regulatory organization such as TPRI in East or CSP in West Africa (Table 7). There are important differences among countries however; in Tanzania all herbicide brands were authorized (by TPRI) but in The Gambia nine out of ten (90%) were not (by CSP). In Togo, 21 of the 26 brands (hence 81%), Côte d’Ivoire, 42 of the 60 brands (70%), and Benin, 13 of the 21 brands (62%), brands did not feature on the CSP list. The two most important herbicide-providing countries were China and France, with the first exporting 50 individual herbicide brands and the second 17. Among the 50 brands of Chinese origin only three were authorized in the CSP list, hence 94% were not. Twenty-four brands were from unknown origin and unregistered.
Table 7

Number of registered (R) and unregistered (U) herbicidesa available in rural markets of importing countries in sub-Saharan Africab, from different herbicide producing countries

Herbicide producing countries

Herbicide importing countries

Individual products

BJ

BF

TD

CI

ML

NE

GM

TG

TZ

R

U

R

U

R

U

R

U

R

U

R

U

R

U

R

U

R

U

Total

Unregistered

China

1

7

 

1

  

3

15

  

1

  

9

2

15

  

50

47

France

4

 

1

   

9

 

5

3

  

1

 

2

1

  

17

4

India

1

1

    

2

    

2

      

5

3

Burkina Faso

  

5

               

5

0

Côte d’Ivoire

1

      

3

1

         

5

3

Ghana

 

1

     

1

       

3

  

5

5

USA

      

2

 

2

       

1

 

4

0

Tanzania

                

4

 

3

0

Singapore

1

1

             

1

  

2

1

Germany

     

1

          

1

 

2

1

Kenya

                

2

 

2

0

Mali

        

1

1

        

2

1

Nigeria

               

2

  

2

2

Switzerland

      

2

           

2

0

Malaysia

       

1

          

1

1

Israel

  

1

               

1

0

South Africa

       

1

          

1

1

UK

 

1

                

1

1

Zambia

                

1

 

1

0

Unknown

 

2

   

1

 

21

          

16

16

Overall

8

13

7

1

0

2

18

42

9

4

1

2

1

9

5

21

9

0

127

86

aFollowing the TPRI ( 2011), for Tanzania and the CSP ( 2013), for the other countries

bBJ: Benin; BF: Burkina Faso; TD: Chad; CI: Côte d’Ivoire; ML: Mali; NE: Niger; GM: The Gambia; TG: Togo; TZ: Tanzania

4 Discussion

4.1 Weed management in rice: Importance of herbicides

Weeds are perceived by farmers as the most important overarching production constraint in rice in SSA (Diagne et al. 2013). Rice systems in SSA are characterized by diverse weed communities. A recent study in East Africa observed 222 species belonging to 46 plant families, with the Poaceae (39 species) and Cyperaceae (38 species) as the most represented ones (Makokha et al. 2017). In a synopsis of the literature on weed species in rice in Africa, Rodenburg and Johnson (2009) reported the five most important species in uplands to be: Rottboellia cochinchinensis, Digitaria horizontalis, Ageratum conyzoides, Tridax procumbens and Eleusine indica. In hydromorphic environments, the top five comprised A. conyzoides, Panicum laxum, Leersia hexandra, Cyperus rotundus, and D. horizontalis and in the lowlands the most frequent species were Sphenoclea zeylanica, Cyperus difformis, Fimbristylis littoralis, Oryza longistaminata, and Echinochloa colona. The reported species diversity makes weed control a complex task.

Weed control proved to be an important management practice for safeguarding rice yields and therefore is an important contributor to increased food security. Based on the data of the current study, it was estimated that farmers in irrigated rice systems could save 1 t ha−1 of grain following a minimum of two weed control interventions. This range of yield loss reductions obtained by weed management corroborates previous studies on irrigated rice (Haefele et al. 2000; Becker et al. 2003). In the rainfed uplands, the rice yield estimate following two weeding interventions was 1.6 t ha−1, a yield loss reduction of 0.4 t ha−1 compared with the no-intervention reference, while three interventions resulted in an estimated yield of 2.1 t ha−1,a loss reduction of nearly 1 t ha−1. These yield savings are similar to estimates from a previous study on weeding in upland rice, conducted by Ogwuike et al. (2014).

For obtaining yield loss reductions, hand weeding alone proved overall less efficient than hand weeding supplemented by either herbicide application or mechanical weed control technologies. In rainfed lowlands, higher yield loss reductions were obtained when hand weeding was supplemented by mechanical weeding, while supplementary herbicide applications did not further reduce yield losses. Herbicides may be less effective here because of the lack of control over water levels as shown before by Toure et al. (2009). Uncontrolled and therefore untimely floods or droughts may cancel out the effectiveness of herbicide applications (Zimdahl 2007). In irrigated lowlands and rainfed upland environments rice yields benefited from herbicide applications as supplementary technology to hand weeding. The survey data however revealed that in order to fulfil the potential of herbicides, the adoption of the technology as well as application practices of the technology need to be improved.

This farm survey showed that 34% of rice farmers in SSA use herbicides to control weeds, although always combined with hand weeding. Wide variation was observed in herbicide use frequencies across countries, confirming an earlier report by Sheahan and Barrett (2017). The low adoption of herbicides in some countries (e.g. DR Congo, Chad, Rwanda, The Gambia, Madagascar and Ethiopia) may have different underlying reasons. In DR Congo, Chad and Rwanda, herbicide availability on the markets seems limited, while in The Gambia herbicides are widely available but come at above-average prices and a large proportion of farmers control weeds mechanically. In Madagascar, farmers rely heavily on cheap family labor for hand weeding (R. Rabeson, personal communication) and rotary weeders are widely adopted (Rodenburg et al. 2015), which is reflected in a high frequency of mechanical weeding. The poor adoption of herbicides in Ethiopia is more difficult to explain, as the herbicide prices are lower than anywhere else in SSA. In teff, a much more traditional and widely grown cereal in the country, a recent and steep increase in herbicide use was observed (Tamru et al. 2017). Among farmers growing rice, which is a relatively new crop, the awareness of this technology could be lower.

Not only in Ethiopia but also in Mali a recent increase in herbicide use in subsistence cereal production was observed (Haggblade et al. 2017a). For rice there is a scarcity of data on herbicide use, which makes it difficult to compare the herbicide adoption figures of the current study to those of the past. Limited reports available on herbicide use by rice farmers in the past however suggest no noteworthy change. Already in the early nineties, herbicides were used by 42% of the rice farmers across rice growing environments in Côte d’Ivoire (compared with 41% in this study), although again mostly combined with hand weeding (Adesina et al. 1994). In the irrigated rice systems of the Senegal River Valley, herbicide use in the late nineties ranged from 60% in Mauritania, to 100% in Senegal (Haefele et al. 2000). The current study showed that herbicides are more often used by men than by women and this confirms the more general observation made by Sheahan and Barrett (2017), that male-headed households more frequently use modern inputs across SSA. This is probably due to gender differences in access to such inputs (Achandi et al. 2018). Also the higher herbicide use observed in irrigated compared with rainfed rice growing environments seems to be a more general feature, as it was previously observed in India (Rao and Nagamani 2010) and the Philippines (Beltran et al. 2013).

Based on the dominance of postemergence herbicides, or post-emergence application of herbicides, it can be concluded that farmers use herbicides more often as a curative control measure, than as a preventive measure. With an observed 38% of likely cases of wrong application timings, the current survey results also indirectly show the weak level of awareness and knowledge at the farmer level concerning herbicides. This is a persistent problem as it was already observed twenty years ago in the Senegal River Valley where farmers were frequently applying herbicides too late and in sub-optimal doses (Haefele et al. 2002). From the current study, particularly worrisome is the high proportion of potential misuse of the controversial broad-spectrum herbicides glyphosate and paraquat.

4.2 Herbicide products: Availability and use

Based on the weed management literature, before 2009 there were 31 herbicide formulations available and used by rice farmers in SSA (Table 8). The current surveys showed market availability of 17 formulations, while farmers were using 18. Of these 18 herbicide formulations, five were not found in the market survey. Three of these, i.e. propanil + butachlor, paraquat + pendimethalin and trifluralin, were observed at locations where no market study was conducted, i.e. Namulonge (Uganda), Hauts-bassins (Burkina Faso) and Afife (Ghana). The other two herbicide formulations, i.e. propanil + bentazon and metolachlor + terbutryn, were probably either out of stock or obtained at other places or through other ways than the agro-chemical supply shops. Eleven herbicide formulations were established (observed before 2009) and the same number of formulations were new (since 2009), or not reported before. Twenty formulations that were cited in the literature did not feature in any of the surveys.
Table 8

Herbicides (formulations, alphabetical order) reported to be available and used in rice in Africa prior to 2009 [according to Rodenburg and Johnson 2009, Akobundu 1987 and Diallo and Johnson 1997] in comparison with the current market survey in 17 countries (26 locations) and farmer survey in 20 countries (36 sites, 1965 farmers) conducted from 2012 to 2015

Herbicide formulations

Literaturea

Markets

Farms

2,4-D

x

x

x

2,4-D + dichlorprop

x

  

2,4-5-TP

x

  

Bensulfuron

x

x

x

Bentazon

x

  

Bifenox

x

  

Butachlor

x

x

x

Cinosulfuron

x

  

Dymrone

x

  

Fluorodifen

x

  

Glyphosate

x

x

x

MCPA

x

  

Molinate

x

  

Oxadiazon

x

x

x

Paraquat

x

x

x

Pendimethalin

x

x

x

Piperophos

x

  

Piperophos + Cinosulfuron

x

  

Pretilhachlor + Dimethametryne

x

  

Propanil

x

x

x

Propanil + Bentazon

x

 

x

Propanil + Fluorodifen

x

  

Propanil + MCPA

x

  

Propanil + Molinate

x

  

Propanil + Triclopyr

x

x

x

Propanil + Piperophos

x

  

Propanil + Oxadiazon

x

  

Propanil + Thiobencarb

x

x

 

Quinclorac

x

  

Thiobencarb

x

  

Triclopyr

x

  

Bensulfuron + Pretilachlor

 

x

 

Bispyribac

 

x

 

Glyphosate + Oxyfluorfen

 

x

x

Haloxyfop-R

 

x

x

Metolachlor + Terbutryn

  

x

Paraquat + Pendimethalin

  

x

Penoxsulam

 

x

 

Pretilachlor + Pyribenzoxim

 

x

x

Propanil +2,4-D

 

x

x

Propanil + Butachlor

  

x

Trifluralin

  

x

Number

31

17

18

aRodenburg and Johnson (2009), Akobundu (1987) and Diallo and Johnson (1997)

Farmers’ access to information and improved technologies is key to reach the necessary increase in rice production for food security in the region (Haefele et al. 2002). Concerning farmer’s access to information, the high illiteracy rate in SSA is likely to be part of the problem. According to UNESCO (2017), out of the 20 countries surveyed in our study, 11 have an adult literacy rate below 50% and only six have a literacy rate between 70 and 79% (i.e. Cameroon, DR Congo, Ghana, Madagascar, Tanzania and Uganda). A recent study on agricultural technologies used by rice farmers in East Africa, with partly the same respondents as in the current study, confirmed the UNESCO reports on Tanzania, Madagascar and Ethiopia (Achandi et al. 2018). The high illiteracy rates in most of the survey countries could explain the low number of farmers who indicated that they had read the product labels to inform themselves about herbicide use.

With respect to access to herbicides a crucial role should be played by the rural supply markets. The market study showed that currently these markets are dominated by only a hand-full of formulations (albeit under many different herbicide brand names) such as glyphosate and 2,4-D. This confirms recent studies in Mali and Ethiopia, summarized by Haggblade et al. (2017a), in which concomitant to an increasing number of mainly Chinese brands on the African markets, the herbicide prices have dropped. Indeed, as the current market survey showed, these products are sold at very competitive prices, i.e. well below the average herbicide price of US $17 ha−1. More than a decade ago, US $10 ha−1 was the average actual level of expenditure on pesticides (mainly herbicides) in SSA (Oerke and Dehne 2004), and therefore this does seem like a price smallholder farmers may be willing to pay. The frequently postulated and observed complaint by farmers that herbicides are too expensive (e.g. Adesina et al. 1994; Tippe et al. 2017) therefore cannot be generalized. On the contrary, herbicides appear to be generally cheaper than wages, as shown in Ethiopia (Tamru et al. 2017), and in Mali where the cost of applying herbicides was less than half the cost of hand weeding (Haggblade et al. 2017b). The comparison between herbicide and hand weeding costs seem to be country specific however. For rice systems in Senegal, Demont et al. (2009) estimated the costs of hand weeding at 15 € ha−1, which was 25% cheaper than their cost estimate for herbicide application.1

Without good stewardship, the heavy reliance on a limited number of herbicide formulations (glyphosate and 2,4-D) may accelerate the development of herbicide resistant weed ecotypes (Davis and Frisvold 2017). The evolution of glyphosate-resistant weed ecotypes illustrates this (Duke and Powles 2008). Apart from the risks concerning the development of herbicide-resistant weed ecotypes, there are concerns about negative herbicide-related impacts on human health and the environment. A number of herbicides that are used by farmers in rice systems in SSA are controversial in this respect. Most prominent are the concerns over the use of glyphosate (Myers et al. 2016), but also 2,4-D has been critically assessed (Peterson et al. 2016) while paraquat is even officially banned in many countries (Haggblade et al. 2017a). Many of the herbicides currently sold on the market are postulated by Haggblade et al. (2017a) to be counterfeit or at least unregistered in the African countries where they are sold. While results of the current study confirm this (62% of available brands were unauthorized), the study also highlighted important differences among countries in terms of the number of unauthorized herbicides. This in turn points to differences in capacities of countries to monitor and regulate pesticide developments at their markets. It has recently been observed that such regulatory capacities of African countries cannot always keep pace with the influx of new herbicide brands, imported from Asia (Haggblade et al. 2017b; Tamru et al. 2017), and this again raises concerns with respect to health and environmental safety.

4.3 The status quo of herbicides in Africa

This study showed that (1) herbicides are potentially important technologies to reduce yield losses and therefore to contribute to food security, (2) herbicides are commonly (32-72%) used by rice farmers in 13 of the 20 countries covered by the survey, and often at multiple application times during the season but, (3) there is a limited diversity of herbicide formulations in supply shops in rural Africa, (4) the most widely available herbicide formulations are glyphosate and 2,4-D and they are among the cheapest available herbicides, (5) herbicide market availability and prices are reflected by what farmers are using in their rice crop, as the same products are predominantly observed here, (6) herbicides are often applied at the wrong time, and (7) farmers make limited use of formal sources of information for the correct application of herbicides. A clear trade-off was observed between the use of formal sources (extension and product labels) and the consultation of neighboring farmers. Suboptimal market processes and communication flows seem to be the most important impediments to the fulfillment of the potential these technologies hold with respect to their contribution to food security and poverty alleviation in the region. As long as this situation does not change, the status quo of poor chemical weed control in rice in SSA will likely be maintained. Prospects for changes are not bright.

For the promotion of modern technologies and good agricultural practices, well-functioning and accessible extension services are imperative (Emmanuel et al. 2016). However, in reality, extension services in SSA are often understaffed, underequipped and often lack the relevant knowledge on, for instance, weed management (e.g. Schut et al. 2015). Alternative means of information transfer, like programs or applications based on information and communication technologies (ICT) such as computers and smart-phones (e.g. Aker 2011; Saito et al. 2015; Rodenburg et al. 2016) or farmer-to-farmer instruction videos (e.g. Zossou et al. 2012) are promising in this respect.

On the supply side, the developments are not conducive either. Globally, innovations in herbicide formulations have been seemingly non-existent since the 1990s (Duke 2012), resulting in an overall low diversity of herbicide formulations even in industrial, developed countries (Davis and Frisvold 2017). Although some recent developments have been noted that point to a renewed interest in herbicide innovations, potential new formulations will likely be much more expensive than existing ones (Haggblade et al. 2017b). The marginal attainable profits for the agro-industry in rural Africa, also do not attract innovative private investments in this area (Demont et al. 2009) and the regional herbicide market is not likely to expand and diversify with already available herbicide formulations in the near future either. The recent trend of increasing imports of cheap herbicides mainly from China and, to a lesser extent, India (Haggblade et al. 2017b), does not contribute to product diversity and quality. Rural Africa is populated by smallholder farmers with small financial margins, who are often unable or reluctant to invest. They will be attracted by the same cheap herbicides imported from Asia and not be incentivized or able to pay more for a new herbicide for which the efficacy has not been proven to them yet. The status quo of herbicides in rice in SSA, is therefore likely to endure.

5 Conclusions

Herbicides are commonly used (32-72%) technologies in 12 of the surveyed countries, while in eight of the countries i.e. Ethiopia, Chad, Madagascar, The Gambia, Tanzania, DR Congo, Rwanda, and Sierra Leone, adoption is less than 3%. Herbicides are more often used by men than by women and more often in irrigated lowland rice than in rainfed rice growing environments.

Herbicides are never used as stand-alone weed technology but rather as a supplement to hand weeding. Compared with hand weeding only, this supplementary use of herbicides can further reduce yield losses by 0.4 t ha−1 in irrigated lowlands and rainfed uplands. Herbicides could therefore play an important role in reaching food security in the region. Based on surveys and recent trends reported in the literature, we however observe a number of problems regarding the sustainability implications of this technology. The global stagnation in herbicide innovations and the poor diversity in available herbicide formulations on local rural markets result in the dominance of few herbicides being used by rice farmers in SSA. Moreover, the rural herbicide supply markets in SSA are dominated by cheap and unregistered herbicide brands, there is a shortfall in effective national regulatory capacities to monitor environmental and health safety related to these products, and we observed a very low rate of users consulting reliable information sources concerning proper and safe herbicide use. The latter is reflected by a high rate of assumed wrong spraying timings.

Overreliance on a small range of herbicide formulations and the frequent use of these formulations in herbicide brands of unknown quality applied at sub-optimal timings/methods may cause negative impacts on the environment, human health, and the crop, and may accelerate the evolution of herbicide resistant weed ecotypes. All these factors jeopardize the future food security in sub-Saharan Africa. Moving away from this status quo, will require (1) improvements in national pesticide regulation procedures and investments for their effective implementation and monitoring of environmental and health impact, (2) innovations in herbicide formulations as well as other labor-saving weed management strategies that stimulate farmers to diversify their approaches, and (3) both innovations and investments that benefit agricultural service provision and communications with farmers specifically concerning the correct choice and timing of herbicides.

Footnotes

  1. 1.

    Conversion rate on 29 May 2018: 1 Euro = 1.16 USD

Notes

Acknowledgements

This is an output of the CGIAR Research program Global Rice Science Partnership. Financial support for this study was provided by the African Development Bank as part of the project “Support to Agricultural Research for Development of Strategic Crops in Africa”. Survey work in Tanzania, Uganda and Rwanda was financially supported by the German Federal Ministry for Economic Cooperation and Development, commissioned by the Deutsche Gesellschaft für Internationale Zusammenarbeit, through the project “East African Wetlands: Optimizing sustainable production for future food security (WETLANDS)”. We thank Justin Djagba of AfricaRice for generating Fig. 1. We thank all farmers and extension personnel for their participation in the surveys.

Compliance with ethical standards

Conflict of interest statement

The authors declared that they have no conflict of interest.

References

  1. Achandi, E. L., Mujawamariya, G., Agboh-Noameshie, A. R., Gebremariam, S., Rahalivavololona, N., & Rodenburg, J. (2018). Women's access to agricultural technologies in rice production and processing hubs: a comparative analysis of Ethiopia, Madagascar and Tanzania. Journal of Rural Studies, 60, 188–198.  https://doi.org/10.1016/j.jrurstud.2018.03.011.
  2. Adesina, A. A., Johnson, D. E., & Heinrichs, E. A. (1994). Rice pests in the Ivory Coast, West Africa - farmers perceptions and management strategies. International Journal of Pest Management, 40(4), 293–299.CrossRefGoogle Scholar
  3. Aker, J. C. (2011). Dial "A" for agriculture: a review of information and communication technologies for agricultural extension in developing countries. Agricultural Economics, 42(6), 631–647.  https://doi.org/10.1111/j.1574-0862.2011.00545.x.CrossRefGoogle Scholar
  4. Akobundu, I. O. (1987). Weed science in the tropics - principles and practices. Chichester: John Wiley & Sons.Google Scholar
  5. Balasubramanian, V., Sie, M., Hijmans, R. J., & Otsuka, K. (2007). Increasing rice production in sub-Saharan Africa: challenges and opportunities. Advances in Agronomy, 94, 55–133.CrossRefGoogle Scholar
  6. Becker, M., & Johnson, D. E. (2001a). Cropping intensity effects on upland rice yield and sustainability in West Africa. Nutrient Cycling in Agroecosystems, 59(2), 107–117.CrossRefGoogle Scholar
  7. Becker, M., & Johnson, D. E. (2001b). Improved water control and crop management effects on lowland rice productivity in West Africa. Nutrient Cycling in Agroecosystems, 59(2), 119–127.CrossRefGoogle Scholar
  8. Becker, M., Johnson, D. E., Wopereis, M. C. S., & Sow, A. (2003). Rice yield gaps in irrigated systems along an agro-ecological gradient in West Africa. Journal of Plant Nutrition and Soil Science, 166(1), 61–67.CrossRefGoogle Scholar
  9. Beltran, J. C., White, B., Burton, M., Doole, G. J., & Pannell, D. J. (2013). Determinants of herbicide use in rice production in the Philippines. Agricultural Economics, 44(1), 45–55.  https://doi.org/10.1111/j.1574-0862.2012.00631.x.CrossRefGoogle Scholar
  10. Comité Sahélien des Pesticides. (2013). Liste des herbicides autorisés par le CSP, Version de novembre 2013. Retrieved from http://www.insah.org/doc/pdf/pesticides_autorises_40%C3%A8e_session_CSP_mai_2017.pdf
  11. Davis, A. S., & Frisvold, G. B. (2017). Are herbicides a once in a century method of weed control? Pest Management Science, 73(11), 2209–2220.  https://doi.org/10.1002/ps.4643.CrossRefPubMedGoogle Scholar
  12. Demont, M., & Rodenburg, J. (2016). On the interaction between weed and bird damage in rice. Weed Research, 56, 193–197.  https://doi.org/10.1111/wre.12206.CrossRefGoogle Scholar
  13. Demont, M., Rodenburg, J., Diagne, M., & Diallo, S. (2009). Ex ante impact assessment of herbicide resistant rice in the Sahel. Crop Protection, 28(9), 728–736.  https://doi.org/10.1016/j.cropro.2009.05.012.CrossRefGoogle Scholar
  14. Diagne, A., Alia, D. Y., Amovin-Assagba, E., Wopereis, M. C. S., Saito, K., & Nakelse, T. (2013). Farmer perceptions of the biophysical constraints to rice production in sub-Saharan Africa, and potential impact of research. In M. C. S. Wopereis, D. E. Johnson, N. Ahmadi, E. Tollens, & A. Jalloh (Eds.), Realizing Africa's rice promise (pp. 46–68). Oxfordshire: CAB International.CrossRefGoogle Scholar
  15. Diallo, S., & Johnson, D. E. (1997). Les adventices du riz irrigué au Sahel et leur contrôle. Paper presented at the Irrigated rice in the Sahel: prospects for sustainable development, Dakar.Google Scholar
  16. Duke, S. O. (2012). Why have no new herbicide modes of action appeared in recent years? Pest Management Science, 68(4), 505–512.  https://doi.org/10.1002/ps.2333.CrossRefPubMedGoogle Scholar
  17. Duke, S. O., & Powles, S. B. (2008). Glyphosate: a once-in-a-century herbicide. Pest Management Science, 64(4), 319–325.CrossRefPubMedGoogle Scholar
  18. Emmanuel, D., Owusu-Sekyere, E., Owusu, V., & Jordaan, H. (2016). Impact of agricultural extension service on adoption of chemical fertilizer: implications for rice productivity and development in Ghana. Njas-Wageningen Journal of Life Sciences, 79, 41–49.  https://doi.org/10.1016/j.njas.2016.10.002.CrossRefGoogle Scholar
  19. Gianessi, L. P. (2013). The increasing importance of herbicides in worldwide crop production. Pest Management Science, 69(10), 1099–1105.  https://doi.org/10.1002/ps.3598.CrossRefPubMedGoogle Scholar
  20. Gongotchame, S., Dieng, I., Ahouanton, K., Johnson, J. M., Alognon, A. D., Tanaka, A., Atta, S., & Saito, K. (2014). Participatory evaluation of mechanical weeders in lowland rice production systems in Benin. Crop Protection, 61, 32–37.  https://doi.org/10.1016/j.cropro.2014.03.009.CrossRefGoogle Scholar
  21. Haefele, S. M., Johnson, D. E., Diallo, S., Wopereis, M. C. S., & Janin, I. (2000). Improved soil fertility and weed management is profitable for irrigated rice farmers in Sahelian West Africa. Field Crops Research, 66(2), 101–113.  https://doi.org/10.1016/s0378-4290(00)00066-6.CrossRefGoogle Scholar
  22. Haefele, S., Wopereis, M. C. S., & Donovan, C. (2002). Farmers' perceptions, practices and performance in a Sahelian irrigated rice scheme. Experimental Agriculture, 38, 197–210.CrossRefGoogle Scholar
  23. Haggblade, S., Minten, B., Pray, C., Reardon, T., & Zilberman, D. (2017a). The herbicide revolution in developing countries: patterns, causes, and implications. European Journal of Development Research, 29(3), 533–559.  https://doi.org/10.1057/s41287-017-0090-7.CrossRefGoogle Scholar
  24. Haggblade, S., Smale, M., Kergna, A., Theriault, V., & Assima, A. (2017b). Causes and consequences of increasing herbicide use in Mali. European Journal of Development Research, 29(3), 648–674.  https://doi.org/10.1057/s41287-017-0087-2.CrossRefGoogle Scholar
  25. HarvestChoice. (2010). Agro-ecological zones of sub-Saharan Africa. Retrieved 12-09-2017, from International Food Policy Research Institute University of Minnesota http://harvestchoice.org/node/8853.
  26. Heinrichs, E. A., Sy, A. A., Akator, S. K., & Oyediran, I. (1997). Seasonal occurrence of rice yellow mottle virus in lowland rice in cote d'Ivoire. International Journal of Pest Management, 43(4), 291–297.CrossRefGoogle Scholar
  27. Johnson, J. M., Rodenburg, J., Tanaka, A., Senthilkumar, K., Ahouanton, K., Dieng, I., et al. (2018). Farmers' perceptions on mechanical weeders for rice production in sub-Saharan Africa. Experimental Agriculture, 1–15.  https://doi.org/10.1017/S001447971700059X.
  28. Makokha, D. W., Irakiza, R., Malombe, I., Le Bourgeois, T., & Rodenburg, J. (2017). Dualistic roles and management of non-cultivated plants in lowland rice systems of East Africa. South African Journal of Botany, 108, 321–330.  https://doi.org/10.1016/j.sajb.2016.08.011
  29. Myers, J. P., Antoniou, M. N., Blumberg, B., Carroll, L., Colborn, T., Everett, L. G., Hansen, M., Landrigan, P. J., Lanphear, B. P., Mesnage, R., Vandenberg, L. N., vom Saal, F. S., Welshons, W. V., & Benbrook, C. M. (2016). Concerns over use of glyphosate-based herbicides and risks associated with exposures: a consensus statement. Environmental Health, 15, 19.  https://doi.org/10.1186/s12940-016-0117-0.CrossRefPubMedPubMedCentralGoogle Scholar
  30. Nhamo, N., Rodenburg, J., Zenna, N., Makombe, G., & Luzi-Kihupi, A. (2014). Narrowing the rice yield gap in east and southern Africa: using and adapting existing technologies. Agricultural Systems, 131, 45–55.  https://doi.org/10.1016/j.agsy.2014.08.003.CrossRefGoogle Scholar
  31. Niang, A., Becker, M., Ewert, F., Dieng, I., Gaiser, T., Tanaka, A., Senthilkumar, K., Rodenburg, J., Johnson, J. M., Akakpo, C., Segda, Z., Gbakatchetche, H., Jaiteh, F., Bam, R. K., Dogbe, W., Keita, S., Kamissoko, N., Mossi, I. M., Bakare, O. S., Cissé, M., Baggie, I., Ablede, K. A., & Saito, K. (2017). Variability and determinants of yields in rice production systems of West Africa. Field Crops Research, 207, 1–12.  https://doi.org/10.1016/j.fcr.2017.02.014.CrossRefGoogle Scholar
  32. Oerke, E.-C., & Dehne, H.-W. (2004). Safeguarding production-losses in major crops and the role of crop protection. Crop Protection, 23(4), 275–285.CrossRefGoogle Scholar
  33. Ogwuike, P., Rodenburg, J., Diagne, A., Agboh-Noameshie, R., & Amovin-Assagba, E. (2014). Weed management in upland rice in sub-Saharan Africa: impact on labor and crop productivity. Food Security, 6, 327–337.  https://doi.org/10.1007/s12571-014-0351-7.CrossRefGoogle Scholar
  34. Peterson, M. A., McMaster, S. A., Riechers, D. E., Skelton, J., & Stahlman, P. W. (2016). 2,4-D past, present, and future: a review. Weed Technology, 30(2), 303–345.  https://doi.org/10.1614/wt-d-15-00131.1.CrossRefGoogle Scholar
  35. Rao, A. N., & Nagamani, A. (2010). Integrated weed management in India–revisited. Indian Journal of Weed Science, 42, 123–135.Google Scholar
  36. R-Core-Team. (2017). R: A language and environment for statistical computing. Vienna, Austria: R Foundation for Statistical Computing Retrieved from https://www.r-project.org/.
  37. Rodenburg, J., & Johnson, D. E. (2009). Weed management in rice-based cropping systems in Africa. Advances in Agronomy, 103, 149–218.  https://doi.org/10.1016/s0065-2113(09)03004-1.
  38. Rodenburg, J., Saito, K., Irakiza, R., Makokha, D. W., Onyuka, E. A., & Senthilkumar, K. (2015). Labor-saving weed technologies for lowland rice farmers in sub-Saharan Africa. Weed Technology, 29(4), 751–757.  https://doi.org/10.1614/WT-D-15-00016.1.CrossRefGoogle Scholar
  39. Rodenburg, J., Le Bourgeois, T., Grard, P., Carara, A., Irakiza, R., Makokha, D. W., … Marnotte, P. (2016). Electronic support tools for identification and management of rice weeds in Africa for better-informed agricultural change agents. Cahiers Agricultures, 25(1), 15006.  https://doi.org/10.1051/cagri/2016004.
  40. Saito, K., Diack, S., Dieng, I., & N'Diaye, M. K. (2015). On-farm testing of a nutrient management decision-support tool for rice in the Senegal River valley. Computers and Electronics in Agriculture, 116, 36–44.  https://doi.org/10.1016/j.compag.2015.06.008.CrossRefGoogle Scholar
  41. Schut, M., Rodenburg, J., Klerkx, L., Kayeke, J., van Ast, A., & Bastiaans, L. (2015). RAAIS: rapid appraisal of agricultural innovation systems (part II). Integrated analysis of parasitic weed problems in rice in Tanzania. Agricultural Systems, 132(0), 12–24.  https://doi.org/10.1016/j.agsy.2014.09.004.CrossRefGoogle Scholar
  42. Seck, P. A., Diagne, A., Mohanty, S., & Wopereis, M. C. S. (2012). Crops that feed the world 7: rice. Food Security, 4(1), 7–24.  https://doi.org/10.1007/s12571-012-0168-1.CrossRefGoogle Scholar
  43. Sheahan, M., & Barrett, C. B. (2017). Ten striking facts about agricultural input use in sub-Saharan Africa. Food Policy, 67, 12–25.  https://doi.org/10.1016/j.foodpol.2016.09.010.CrossRefPubMedPubMedCentralGoogle Scholar
  44. Tamru, S., Minten, B., Alemu, D., & Bachewe, F. (2017). The rapid expansion of herbicide use in smallholder agriculture in Ethiopia: patterns, drivers, and implications. The European Journal of Development Research, 29(3), 628–647.  https://doi.org/10.1057/s41287-017-0076-5.CrossRefGoogle Scholar
  45. Tanaka, A., Johnson, J.-M., Senthilkumar, K., Akakpo, C., Segda, Z., Yameogo, L. P., Bassoro, I., Lamare, D. M., Allarangaye, M. D., Gbakatchetche, H., Bayuh, B. A., Jaiteh, F., Bam, R. K., Dogbe, W., Sékou, K., Rabeson, R., Rakotoarisoa, N. M., Kamissoko, N., Mossi, I. M., Bakare, O. S., Mabone, F. L., Gasore, E. R., Baggie, I., Kajiru, G. J., Mghase, J., Ablede, K. A., Nanfumba, D., & Saito, K. (2017). On-farm rice yield and its association with biophysical factors in sub-Saharan Africa. European Journal of Agronomy, 85, 1–11.  https://doi.org/10.1016/j.eja.2016.12.010.CrossRefGoogle Scholar
  46. Tippe, D. E., Rodenburg, J., Schut, M., van Ast, A., Kayeke, J., & Bastiaans, L. (2017). Farmers’ knowledge, use and preferences of parasitic weed management strategies in rain-fed rice production systems. Crop Protection, 99, 93–107.  https://doi.org/10.1016/j.cropro.2017.05.007.CrossRefGoogle Scholar
  47. Touré, A., Becker, M., Johnson, D. E., Koné, B., Kossou, D. K., & Kiepe, P. (2009). Response of lowland rice to agronomic management under different hydrological regimes in an inland valley of Ivory Coast. Field Crops Research, 114, 304–310.  https://doi.org/10.1016/j.fcr.2009.08.015.CrossRefGoogle Scholar
  48. Tropical Pesticide Research Institute. (2011). List of registered pesticides in Tanzania. Retrieved from http://tpri.or.tz/news/Pesticides_Gazette_2011.pdf
  49. UNESCO. (2017). Fact Sheet No. 45, September 2017 FS/2017/LIT/45. UNESCO Institute for Statistics. http://uis.unesco.org/sites/default/files/documents/fs45-literacy-rates-continue-rise-generation-to-next-en-2017_0.pdf
  50. Waddington, S. R., Li, X. Y., Dixon, J., Hyman, G., & De Vicente, M. C. (2010). Getting the focus right: production constraints for six major food crops in Asian and African farming systems. Food Security, 2(1), 27–48.  https://doi.org/10.1007/s12571-010-0053-8.CrossRefGoogle Scholar
  51. Zimdahl, R. L. (2007). Fundamentals of weed science (Vol. 3rd). London: Academic Press.Google Scholar
  52. Zossou, E., Van Mele, P., Wanvoeke, J., & Lebailly, P. (2012). Participatory impact assesment of rice parboiling videos with women in Benin. Experimental Agriculture, 48(3), 438–447.  https://doi.org/10.1017/s0014479712000117.CrossRefGoogle Scholar

Copyright information

© The Author(s) 2019

Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made.

Authors and Affiliations

  • Jonne Rodenburg
    • 1
    • 2
    Email author
  • Jean-Martial Johnson
    • 2
    • 3
  • Ibnou Dieng
    • 2
  • Kalimuthu Senthilkumar
    • 4
  • Elke Vandamme
    • 5
  • Cyriaque Akakpo
    • 6
  • Moundibaye Dastre Allarangaye
    • 7
  • Idriss Baggie
    • 8
  • Samuel Oladele Bakare
    • 9
  • Ralph Kwame Bam
    • 10
  • Ibrahim Bassoro
    • 11
  • Bayuh Belay Abera
    • 12
  • Madiama Cisse
    • 13
  • Wilson Dogbe
    • 14
  • Henri Gbakatchétché
    • 15
  • Famara Jaiteh
    • 16
  • Geophrey Jasper Kajiru
    • 17
  • Alain Kalisa
    • 18
  • Nianankoro Kamissoko
    • 19
  • Keita Sékou
    • 20
  • Ahouanton Kokou
    • 2
  • Delphine Mapiemfu-Lamare
    • 11
  • Fanny Mabone Lunze
    • 21
  • Jerome Mghase
    • 22
  • Illiassou Mossi Maïga
    • 23
  • David Nanfumba
    • 24
  • Abibou Niang
    • 25
  • Raymond Rabeson
    • 26
  • Zacharie Segda
    • 27
    • 28
  • Fitta Silas Sillo
    • 29
  • Atsuko Tanaka
    • 30
  • Kazuki Saito
    • 2
  1. 1.Natural Resources Institute (NRI)University of GreenwichKentUK
  2. 2.Africa Rice Center (AfricaRice)Bouake 01Côte d’Ivoire
  3. 3.Institute of Crop Science and Resource Conservation (INRES)University of BonnBonnGermany
  4. 4.Africa Rice Center (AfricaRice)AntananarivoMadagascar
  5. 5.International Potato Center (CIP)KigaliRwanda
  6. 6.Institut National des Recherches Agricoles du Bénin (INRAB)BohiconBenin
  7. 7.Institut Tchadien de Recherche Agronomique pour le Développement (ITRAD)N’DjaménaChad
  8. 8.Rokupr Agricultural Research Centre (RARC)Sierra Leone Agricultural Research Institute (SLARI)Tower HillSierra Leone
  9. 9.National Cereals Research Institute (NCRI)BadeggiNigeria
  10. 10.CSIR-Crops Research InstituteKumasiGhana
  11. 11.Institut de Recherche Agricole pour le Développement (IRAD)YaoundeCameroon
  12. 12.Ethiopian Institute of Agricultural Research (EIAR)Bahir DarEthiopia
  13. 13.Centre de Recherches Agricoles (CRA) de Saint-LouisInstitut Sénégalais de Recherches Agricoles (ISRA)Saint-LouisSenegal
  14. 14.CSIR Savanna Agricultural Research Institute (SARI)TamaleGhana
  15. 15.Centre National de Recherche Agronomique (CNRA)ManCôte d’Ivoire
  16. 16.National Agricultural Research Institute (NARI)BrikamaGambia
  17. 17.Department of Research and DevelopmentMinistry of AgricultureDar es SalaamTanzania
  18. 18.Rwanda Agricultural Board (RAB)KigaliRwanda
  19. 19.Institute d’Economie Rurale (IER), CRRANionoMali
  20. 20.Institut de Recherche Agronomique de Guinée (IRAG)ConakryGuinea
  21. 21.Institut National pour l’Etude et la Recherche Agronomiques (INERA)KinshasaDemocratic Republic of Congo
  22. 22.Kilombero Agricultural Research and Training Institute (KATRIN)IfakaraTanzania
  23. 23.Institut National de la Recherche Agronomique du Niger (INRAN)NiameyNiger
  24. 24.National Agricultural Research Organization (NARO)MbaleUganda
  25. 25.Olam International Ltd. In GabonLibrevilleGabon
  26. 26.Centre National de Recherche Appliquée au Développement Rural (FOFIFA)AntananarivoMadagascar
  27. 27.Centre National de la Recherche Scientifique et TechnologiqueInstitut de l’Environnement et de Recherches Agricoles (CNRST/INERA)Bobo DioulassoBurkina Faso
  28. 28.Bureau National des Sols (BUNASOLS)OuagadougouBurkina Faso
  29. 29.Africa Rice Center (AfricaRice), East and Southern AfricaDar es SalaamTanzania
  30. 30.Japan International Cooperation Agency (JICA), Benin OfficeCotonouBenin

Personalised recommendations