Soil Science Society of America Journal 63:972-976 (1999)
© 1999 Soil Science Society of America
DIVISION S-6-SOIL & WATER MANAGEMENT & CONSERVATION
Shallow Tillage with a Traditional West African Hoe to Conserve Soil Water
W.A. Paynea
a Oregon State Univ., Columbia Basin Agricultural Research Center, P.O. Box 390, Pendleton, OR 97886 USA
william.payne{at}orst.edu
 |
ABSTRACT
|
|---|
Evaporation (E) from soil constitutes a large proportion of evapotranspiration (ET) of pearl millet [Pennisetum glaucum (L.) R. Br.] fields in West Africa. Reduced E would increase crop water supply and reduce the risk of resowing. We hypothesized that tillage after rain events with the hilaire, a traditional, shallow-cultivating hoe that pulverizes and darkens the soil surface, could conserve soil water and increase pearl millet yield. Experiments were conducted using bare and cropped plots. Tillage after rain (+ or -) was the only treatment for bare plots. Treatments for cropped plots were tillage after rain with the hilaire and fertilizer addition (12 kg ha-1 N and 12 kg P ha-1). For bare plots, tillage with the hilaire reduced soil surface reflectance by as much as 
0.2 and increased soil temperature at 0.05 m by 12°C. Tillage increased soil water storage in the upper 2.4 m of bare plots by up to 47 mm. For cropped plots, tillage increased soil water storage in the upper 1.4 m by up to 32 mm, and it increased grain yield by 68% in 1991 and 70% in 1992. Fertilizer addition increased yield by 21% in 1991 and 116% in 1992. Tillage reduced ET in 1992 from 417 to 372 mm and increased water-use efficiency (WUEET) from 0.99 to 1.91 kg ha-1 mm-1. Fertilizer addition increased WUEET from 0.95 to 1.94 kg ha-1 mm-1. Practical exploitation of the hilaire's effect upon evaporation is not expected because it is a hand-operated tool. However, the study demonstrates principles upon which an animal-drawn implement might be designed.
Abbreviations: E, evaporation from soil • ET, evapotranspiration (evaporation from soil + crop transpiration) • WUEET, water-use efficiency (grain yield/ET)
|
INTRODUCTION
|
|---|
PEARL MILLET LEAF AREA INDICES are typically 
0.5 during early growth stages in semi-arid West Africa, causing transpiration to constitute a relatively small fraction of ET (Daamen, 1997). The probability of dry spells 
10 d during this growth period is high (Sivakumar, 1992), and crop water supply is often exhausted, necessitating resowing after the next sufficiently large rain event. Delayed sowing is generally associated with yield decline in pearl millet (Reddy and Visser et al., 1993) and other photosensitive cereals (Andrews, 1973). Any reduction of E during this and subsequent periods would increase water supply for crop growth and reduce the risk of resowing.
Practical methods of reducing evaporation from soils to conserve water are lacking in West Africa. The use of organic mulch during the growing season is desirable but not practicable because most crop residues are fed to livestock or used for building materials during the long dry season. The use of contoured plastic mulch has been shown to be technologically feasible (Zaongo et al., 1994), but it is also unlikely to be practiced widely because such materials are too expensive or generally unavailable in most of West Africa. Mulching at the end of the growing season and bare fallowing have proven ineffectual for water conservation, because conserved water rapidly drains through the soil profile during the dry season (Payne et al., 1990a). A method is needed, therefore, to control evaporation during the growing season that does not rely upon the use of valuable crop residues or expensive synthetics.
Hillel (1982) states that evaporative flux can be modified in three basic ways: (i) by controlling energy supply to the site of evaporation; e.g., modify soil albedo through color; (ii) by reducing the potential gradient driving water upward through the profile; e.g., warm the surface so as to set up a downward-acting thermal gradient; and (iii) by deceasing the conductivity of the profile, particularly of the surface zone; e.g., till the soil. The best choice depends, among other things, upon whether one intends to regulate the first stage of evaporation, during which meteorological conditions dominate the rate of evaporation, or the second stage, during which soil hydraulic properties become limiting.
In sandy soils such as those typically sown to millet in Niger, Mali, and Senegal, water is conducted very rapidly at high water content and very slowly at low water content (Hartman and Gandah, 1982; Payne et al., 1991). Furthermore, atmospheric evaporative demand can be great following the convective storms that typify rain events in the early part of the growing season (Daamen, 1997). Theoretically, it would be most advantageous in these conditions to interrupt soil water evaporation during the first stage.
An approach discussed by Hillel (1982) for controlling evaporation during the first stage is to induce a temporarily higher evaporation rate so as to rapidly dessicate the surface, thus hastening the end of the first stage and using the hysteresis effect to help arrest or retard subsequent outflow. Heating the surface by darkening its color has been suggested, but not tested. Another approach Hillel (1982) discusses is the use of shallow cultivation to pulverize the soil surface and cause the loosened layer to dry faster and more completely. This could over a period of time help conserve soil moisture.
The hilaire is a shallow cultivating, traditional hoe that has been used for centuries on sandy soils of West Africa to control weeds (Fig. 1)
. It is pushed and then pulled by the user such that the blade cuts the roots of weeds at 4 to 5 cm below the soil surface. The affected surface is pulverized and loosened. Furthermore, the color of the soil's surface becomes darker because the underlying soil layer has greater organic matter pigmentation. In effect, the tillage with the hilaire brings about each of the changes described by Hillel (1982) that could reduce E from bare soils.
It has been shown repeatedly that pearl millet has relatively low water-use efficiency (WUEET; kg grain mm-1 ET), and that moderate additions of fertilizer can dramatically increase both yield and WUEET (Cissé and Vachaud, 1988; Klaij and Vachaud, 1992; Payne et al., 1996; Payne, 1997). Any water conserved by reduced E would be much more efficiently used for crop growth when fertilizer is added.
The purpose of this study was to measure the effects of tillage by the hilaire upon soil water conservation, and to test whether these putative effects, with and without the addition of fertilizer, increase yield and WUEET of pearl millet.
|
Materials and methods
|
|---|
Field experiments were carried out in 1991 and 1992 at the ICRISAT Sahelien Center at Sadoré, Niger, and in 1991 at the INRAN (Institut National de Recherches Agronomiques du Niger) research station in Ouallam, Niger. Rainfall is mono-modally distributed in this region of Africa, with rains generally beginning in May or June and ending in August or September. Sadoré has a mean annual rainfall of 580 mm, whereas Ouallam's mean annual rainfall is only 450 mm. Ouallam is considered to be a more marginal site because of lower and more erratic rainfall and greater incidence of dust storms. The experiment was not repeated at Ouallam in 1992 because of political instability.
Surface soil horizons at the Sahelian Center typically have a sandy loam to loamy sand texture, a pH in water (1:1) of 5.5, a cation-exchange capacity of 0.9 cmol kg-1, and an organic matter content of 3 g kg-1. In the U.S. system of soil taxonomy, they are classified as sandy, silicious, isohypothermic Psammentic Kandiustalfs. Soils at Ouallam are similar physically but haven't been classified in the U.S. system.
At each site, 10- by 10-m bare and cropped plots were laid out in a randomized complete block design. For bare plots, tillage was the only treatment, with four replicates for each treatment. In the tilled plots, the hilaire was used to scarify the soil after each rain event 4 mm in 1991, and 15 mm in 1992. Tillage was effected as soon as rains stopped if the rains occurred during the day, and at first light if rains came at night. In the nontilled plots, weeds were killed with a combination of atrazine (6-chloro-N-ethyl-N'-isopropyl-1,3,5-triazine-2,4-diamine) and fluometuron {N, N-dimethyl-N'-[3-(trifluoromethyl) phenyl]urea}.
In the cropped plots, tillage and fertilizer were treatments, with four replicates for each treatment. Pearl millet was planted by hand after the first rain 20 mm at the traditional planting density of one hill m-2. At approximately 20 d after sowing, millet was thinned to three plants per hill. Fertilizer treatments were zero fertilizer addition and, in the fertilized plots, 12 kg ha-1 N (as urea) and 12 kg P ha-1 of single super phosphate. These represent low-input conditions. Phosphorus was broadcast by hand before planting, and urea was broadcast by hand near flowering. For untilled plots in 1991, atrazine was used to control weeds. Because of suspected damage to plants, however, in 1992 plots were weeded with the hilaire at 20 d after sowing, as is customary within the region. Weeds were hand-pulled thereafter.
Each cropped and bare plot replicate had one neutron probe access tube installed to a depth of 2.4 m. Neutron probe readings were taken from four to seven times in 1991. In 1992, probe readings were taken approximately weekly in order to correct ET estimates for root zone drainage (Klaij and Vachaud, 1992). Manual rainfall gauges were installed at the experimental plots. Evapotranspiration was calculated using the water balance approach described by Payne (1997).
One tilled and one untilled bare plot were each instrumented with solarimeters1
(Delta-T Devices Ltd, Cambridge, UK; Szeicz et al., 1964) facing up and down, and with thermocouples at a depth of 0.05 m. Solarimeter and thermocouple data were recorded hourly with a data logger.
Statistical analyses for yield, ET, WUEET, and soil water storage were made using SYSTAT vers. 7.0 (SPSS, Chicago). Soil water storage was calculated to a depth of 1.40 m for the cropped plots, because this is the approximate depth of the root zone (Payne et al., 1990b) and to a depth of 2.4 m (the depth of access tubes) in bare plots.
|
Results and discussion
|
|---|
Dates and amounts of rain events are presented in Fig. 2
. Total rainfall was 550 mm at Sadoré in 1991, 499 mm at Ouallam in 1991, and 483 mm at Sadoré in 1992. There were fewer rain events at Ouallam than at Sadoré, and they were often accompanied by violent dust storms that buried seedlings. This necessitated repeated replanting where seedlings had been buried within plots, and it resulted in a poor and variable stand.
Soil reflectance was lower in the tilled plot than in the untilled plot (Fig. 3)
. Near noon, reflectance was reduced by as much as 0.2 in the tilled plot. Corresponding greater energy absorption was confirmed by soil temperature at 0.05 m, which was as much as 12°C higher in the tilled plot than in the untilled plot. Greater energy absorbance is consistent with Hillel's (1982) description of limiting cumulative evaporation from the soil surface by inducing a temporarily higher evaporation rate.
View larger version (16K):
[in this window]
[in a new window]
|
Fig. 3 Soil surface reflectance and soil temperature at 0.05 m as affected by shallow tillage with the hilaire
|
|
Soil water storage was greater in bare tilled plots at both sites (Fig. 4)
. Tilled plots contained 15 to 20 mm more water than untilled plots in the early part of the 1991 season at Sadoré. Most of this additional water was near the surface, where seedling roots would benefit. In 1991, tilled plots had 47 mm more water than untilled plots on Day of Year 230, which was 80 d after sowing. Thereafter, because of heavy rains (Fig. 2) and greater hydraulic conductivity associated with high soil water content, tilled plots rapidly drained. By the beginning of the 1992 season, mean water storage in the upper 2.4 m of soil was only 18 mm greater in tilled plots than in nontilled plots. A similar trend of rapid drainage of conserved water was reported by Payne et al. (1990a) for bare fallowed plots.
View larger version (18K):
[in this window]
[in a new window]
|
Fig. 4 Soil water storage in the upper 2.6 mm of bare plots as affected by tillage with the hilaire. Points represent mean values from four plots. Bars represent ±1 SE
|
|
In 1992, when plots at Sadoré were tilled less often, the effects of tillage with the hilaire upon soil water conservation in the bare plots were not nearly so great (Fig. 4). Tilled plots had a maximum of 26 mm more water stored than untilled plots. At Ouallam in 1991, tilled bare plots had a maximum of 27 mm more water stored than untilled plots.
Water stored in the root zone (1.40 m) was greater in tilled cropped plots than nontilled plots for both years at Sadoré (Fig. 5)
. Similarly to bare plots, more water was conserved in 1991 than in 1992. There was a maximum of 32 mm more water stored in tilled plots compared with untilled plots in 1991, and 22 mm more in 1992. At Ouallam, however, there were no differences in soil water storage between tilled and untilled cropped plots, presumably due to the very great stand variability caused by dust storms.
View larger version (21K):
[in this window]
[in a new window]
|
Fig. 5 Soil water storage in the upper 1.4 mm of cropped plots as affected by tillage with the hilaire. Points represent mean values from four plots. Bars represent ±1 SE
|
|
Grain and dry matter yield of pearl millet increased because of tillage with the hilaire and the addition of fertilizer in both years at Sadoré (Table 1)
. These effects were additive, as there was no interactive effect between tillage and fertilizer in either year (Table 2)
.
View this table:
[in this window]
[in a new window]
|
Table 1 Grain and dry matter (DM) yield of pearl millet as affected by use of the hilaire and fertilizer application at Sadoré, Niger. Total crop evapotranspiration (ET) and water-use efficiency (WUEET; i.e., grain yield/ET) are shown for 1992 only
|
|
View this table:
[in this window]
[in a new window]
|
Table 2 ANOVA results for grain and dry matter yield of pearl millet as affected by tillage with the hilaire and fertilizer application in 1991 and 1992 at Sadoré, Niger. ANOVA for the effects of tillage and fertilizer application on evapotranspiration (ET) and water-use efficiency (WUEET) are shown for 1992
|
|
Because of the poor and variable stand establishment at Ouallam, yield was affected neither by tillage nor by fertilizer addition (data not shown). The only discernable positive effect upon yield was leeward distance from the experiment's upwind boundary, which was the most exposed to the devastating effects of wind and seedling burial. This demonstrates that fertilizer addition and improved soil water conservation alone will be insufficient to increase pearl millet yield at marginal sites that are prone to dust storms. The effects of dust storms and wind erosion on pearl millet growth were reported by Michels et al. (1993).
Water balance calculations indicated that ET was 45 mm less in tilled plots than in untilled plots (Table 1). Addition of fertilizer increased mean pearl millet ET by 13 mm, but the this was not statistically significant at P < 0.05. Small increases in pearl millet ET that are due to fertilizer addition have been reported elsewhere (Payne et al., 1996; Payne, 1997).
Because yield increased due to tillage with the hilaire, while ET decreased, WUEET in tilled plots doubled (Table 1). Fertilizer addition increased yield much more than it did ET, also resulting in a near doubling of WUEET. Pearl millet grown with neither tillage nor fertilizer had a mean WUEET value of 0.48 kg ha-1 mm-1, whereas plants that received both tillage and fertilizer had a mean WUEET value of 2.39 kg ha-1 mm-1. Thus, the combination of these two factors increased WUEET nearly fivefold. Nonetheless, 2.39 kg ha-1 mm-1 is fairly low compared to pearl millet WUEET values obtained by Payne (1997) in the Sahel that were as high as 7.8 kg ha-1 mm-1. Higher values were obtained with more intensive management, including advanced varieties, increased plant populations, and greater fertilizer addition.
It would be unrealistic to expect subsistence farmers to till entire fields with the hand-operated hilaire after each rain event. Indeed, labor shortage constitutes one of the biggest constraints to pearl millet production in West Africa (Kutzbach et al., 1990; Williams, 1997). Our results do demonstrate, however, that timely shallow tillage after a rain event can conserve water and increase yield dramatically. In order to render this principle useful to farmers, an animal-drawn implement would need to be designed that reproduces the hilaire's effect.
In collaboration with ICRISAT, researchers at the University of Hohenheim (Germany) developed a donkey-drawn implement, termed the hata, which has three sweeps fashioned after the shape of the hilaire blade (Kutzbach et al., 1990). The hata can substantially reduce the time and labor associated with weeding compared with the hand-operated hilaire. However, the sweeps move only forward, whereas the pulverizing effect of the hilaire is dependent upon a back-and-forth motion. It seems doubtful, therefore, that tillage with the hata will have the same effect upon soil water conservation as tillage with the hilaire. An additional implement might be added behind the sweeps to effect this pulverization, similar to the rod weeder used to develop a dust mulch in summer-fallowed fields of the U.S. Pacific Northwest. If such an implement could be adapted to donkey traction, it would hold much potential to conserve soil water, improve WUEET, and increase food production in semi-arid West Africa.
|
Conclusions
|
|---|
Use of the hilaire in sandy soils of semi-arid West Africa after rain events reduces evaporation from the soil surface, reduces crop ET, conserves soil moisture, and increases yield and WUEET. Since it is a hand-operated tool in a region of labor shortage, it is doubtful whether this effect can be exploited practically. The research does demonstrate, however, the principles upon which a practical, animal-drawn implement can be designed.Kutzback Betker Klaij 1993; Reddy Visser 1993
|
ACKNOWLEDGMENTS
|
|---|
The author thanks Mara Boubacar and Moustapha Amadou for invaluable technical assistance, and researchers at INRAN for collaborative support. Thanks also to Jim Gardiner for taking preliminary field measurements several years ago under difficult working conditions.
|
NOTES
|
|---|
1 Mention of trademark names does not constitute an endorsement. 
Received for publication February 9, 1998.
|
REFERENCES
|
|---|
- Andrews D.J. Effects of date of sowing on photosensitive Nigerian sorghums. Exp. Agric. 1973;9:337-346.
- Cissé L., Vachaud G. Influence d'apports de matière organique sur la culture de mil et d'arachide sur un sol sableux du Nord-Sénégal: I. Bilans de consommation, production et développement racinaire. Agronomie 1988;8:315-326.
- Daamen C.C. Two source model of surface fluxes for millet fields in Niger. Agricultural For. Meteorol. 1997;83:205-230.
- Hartman R., Gandah M. Hydrodynamic characterization of a sand dune soil in Niger. Meded. Fac. Landbouwwet. Rijksuniv. Gent. 1982;47:1195-1204.
- Hillel D. Introduction to soil physics. New York: Academic Press, 1982.
- Klaij M.C., Vachaud G. Seasonal water balance of a sandy soil in Niger cropped with pearl millet, based on profile moisture measurements. Agric. Water Manage. 1992;21:313-330.
- Kutzback, H.D., J. Betker, and M. Klaij. 1993. Appropriate mechanization of field work and transportation. p. 312–338. In Sonderforschungsbereich 308. Standortgemäße Landwirtschaft in den Tropen mit Forschungsschwerpunkt Westafricka. Universität Hohenheim, Stuttgart, Germany.
- Michels K., Sivakumar M.V.K., Allison B.E. Wind erosion in the Southern Sahelian zone and induced constraints to pearl millet production. Agric. For. Meteorol. 1993;67:65-77.
- Payne W.A. Managing yield and water use of pearl millet in the Sahel. Agron. J. 1997;89:481-490.[Abstract/Free Full Text]
- Payne, W.A., H. Brück, B. Sattelmacher, S.V.R. Shetty, and C. Renard. 1996. Root growth and soil water extraction of three pearl-millet varieties at three phenological stages. p. 251–259. In Dynamics of roots and nitrogen in cropping systems of the semi-arid tropics. Proc. Int. Worksh., Patancheru, India. 21–25 Nov. 1994. ICRISAT, Patancheru, India.
- Payne W.A., Lascano R.J., Wendt C.W. Physical and hydrological characterization of three sandy millet fields in Niger. In: Sivakumar, et al. , ed. Soil water balance in the Sudano-Sahelian zone. Proc. Int. Worksh., Niamey, Niger, Feb. 1991. IAHS Publ. 199. Wallingford, UK: IAHS Press, Inst. of Hydrology, 1991:199-208.
- Payne W.A., Wendt C.W., Lascano R.J. Bare fallowing on sandy fields of Niger, West Africa. Soil Sci. Soc. Am. J. 1990;54:1079-1084 a.[Abstract/Free Full Text]
- Payne W.A., Wendt C.W., Lascano R.J. Root zone water balances of three low-input millet fields in Niger, West Africa. Agron. J. 1990;82:813-819 b.[Abstract/Free Full Text]
- Reddy K.C., Visser P.L. Planting effects on early versus late pearl millet genotypes in Niger. Exp. Agric. 1993;29:121-129.
- Szeicz G., Monteith J.L., Dos Santos J. A tube solarimeter to measure radiation among plants. J. Appl. Ecol. 1964;1:169-174.
- Sivakumar M.V.K. Empirical analysis of dry spells for agricultural applications in West Africa. J. Clim. 1992;5:532-539.
- Williams T.O. Problems and prospects in the utilization of animal traction in semi-arid west Africa: Evidence from Niger. Soil Tillage Res. 1997;42:295-311.
- Zaongo C.G., Hossner L.R., Wendt C.W. Root distribution, water use, and nutrient uptake of millet and grain sorghum on West African soils. Soil Sci. 1994;157:379-388.
TOP
ABSTRACT
INTRODUCTION
