This article was downloaded by: [Wageningen UR] On: 2 March 2011 Access details: Access Details: [subscription number 932819493] Publisher Taylor & Francis Informa Ltd Registered in England and Wales Registered Number: 1072954 Registered office: Mortimer House, 37- 41 Mortimer Street, London W1T 3JH, UK

Communications in Science and Plant Analysis Publication details, including instructions for authors and subscription information: http://www.informaworld.com/smpp/title~content=t713597241

Soil nutrient status of smallholder farms in Malawi S. S. Snappa a International Crops Research Institute for the Semi-Arid Tropics (ICRISAT), Lilongwe, Malawi

To cite this Article Snapp, S. S.(1998) 'Soil nutrient status of smallholder farms in Malawi', Communications in and Plant Analysis, 29: 17, 2571 — 2588 To link to this Article: DOI: 10.1080/00103629809370135 URL: http://dx.doi.org/10.1080/00103629809370135

PLEASE SCROLL DOWN FOR ARTICLE

Full terms and conditions of use: http://www.informaworld.com/terms-and-conditions-of-access.pdf

This article may be used for research, teaching and private study purposes. Any substantial or systematic reproduction, re-distribution, re-selling, loan or sub-licensing, systematic supply or distribution in any form to anyone is expressly forbidden.

The publisher does not give any warranty express or implied or make any representation that the contents will be complete or accurate or up to date. The accuracy of any instructions, formulae and drug doses should be independently verified with primary sources. The publisher shall not be liable for any loss, actions, claims, proceedings, demand or costs or damages whatsoever or howsoever caused arising directly or indirectly in connection with or arising out of the use of this material. COMMUN. SOIL SCI. PLANT ANAL., 29(17&18), 2571-2588 (1998)

Soil Nutrient Status of Smallholder Farms in Malawi1

S. S. Snapp2

International Crops Research Institute for the Semi-Arid Tropics (ICRISAT), P.O. Box 1096, Lilongwe, Malawi

ABSTRACT

A soil sampling exercise was conducted throughout 60% of the smallholder farm sector of Malawi, a small country located at the base of the Great Rift Valley. Soil samples (1,130) were geo-referenced and analyzed for pH, texture, soil organic carbon (C), phosphorus (P), zinc (Zn), potassium (K), and calcium (Ca) status. Descriptive statistics of soil characteristics were used to evaluate soil fertility for two agricultural districts in Northern Malawi, two agricultural districts in Central Malawi, and one agricultural district in Southern Malawi. Generally were loamy and moderately acid, with "low" to "sufficient" nutrient levels. Over three-quarters of soils sampled Downloaded By: [Wageningen UR] At: 14:24 2 March 2011 had organic C levels which were greater than 0.8%. This indicated that organic C status was adequate in the main to maintain , although much reduced from the non-cultivated state. The organic C data supported observations of widespread nitrogen (N) deficiency in Malawi. High spatial variability of P and Zn values was noted. However, over 60% of soils had a

1Research funded by the Rockefeller Foundation, P.O. Box 30721, Lilongwe 3, Malawi, in cooperation with the Malawi Ministry of Agriculture. 2E-mail address: [email protected].

2571

Copyright © 1998 by Marcel Dekker, Inc. www.dekker.com 2572 SNAPP

P status above the critical value of 15 mg kg-1 which was sufficient for smallholder maize production levels. This finding supported efforts to alter the current country-wide fertilizer recommendation of 45 kg ha-1 phosphate. New findings reported here were location specific edaphic problems: i) widespread moderate soil acidity in Central Malawi, ii) natural regions in Central and Southern Malawi which were low in P and Zn, and iii) natural regions in Southern Malawi were very low in organic C. Research requirements were identified. Suggested priorities for technology development were those appropriate to smallholder farmers, such as combined use of small amounts of inorganic and organic fertilizers to address very nutrient deficient soils.

INTRODUCTION

Soil fertility status is the foundation of cropping system productivity in the smallholder agriculture sector of Southern Africa. A case in point is Malawi. A small country located at the base of the Great Rift Valley, over 85% of the population is engaged in smallholder agriculture. Use of fertilizer inputs in this sector is about 20 kg ha1 (Heisy and Mwangi, 1996). Fertilizer recommendations in Malawi have been broadly written, with limited attention to resource variability and a cash-constrained farming environment (Snapp and Benson, 1995; Kumwenda et al., 1997). Nitrogen (N) has been shown to be generally the most crop limiting nutrient. Phosphorus (P), sulfur (S), and zinc (Zn) responses have also been shown for specific crops and regions in Malawi (Brown and Young, 1966; Chilimba, 1996; Snapp et al., in press). A better understanding of soil nutrient status in Malawi would facilitate development of targeted recommendations. Technologies to maximize nutrient use efficiency and improve return on a scarce resource are urgently needed. Downloaded By: [Wageningen UR] At: 14:24 2 March 2011 Current fertilizer recommendations do not take into account diversity in soil types, farmer resources or yield goals. For example, a blanket fertilizer 1 1 recommendation of 92 kg N ha" N and 45 kg P205 ha" ha is used for hybrid maize (Zea mays L.) production throughout the country, where maize is the dominant crop in Malawi smallholder agriculture (Guide to Agriculture Production, Malawi Ministry of Agriculture and Livestock Development, Lilongwe, Malawi, 1995). Several decades of research in Malawi has addressed maize response to fertilizer (Brown and Young, 1966; Kumwenda et al., 1997). Fertilizer recommendations for maize are in the process of being revised based on fertilizer response trial data. In contrast to fertilizer trials, limited research has addressed soil nutrient status and patterns of micro and macro nutrient deficiency (Chilimba, 1996; Wendt, 1993). Soil morphological characteristics have been studied in the field, and attempts made to relate parent material to chemical status (Brown and Young, 1966; Mandemere and Robertson, 1975). SOIL NUTRIENT STATUS OF SMALLHOLDER FARMS 2573

Soil types in Malawi are in the main , with and also of importance. Soil textures are generally sands and sandy (Kumwenda et al., 1997; Snapp et al., 1988). However, it remains a considerable challenge to characterize soil fertility of smallholder farming systems in Malawi. Extremes in topography confounded with small field size and diversity in crop management history have led to a situation where soil fertility can vary by orders of magnitude across short distances (Snapp and Benson, 1995). Topography in Malawi is characterized by a highly dissected and undulating landscape (Brown and Young, 1966; Young and Brown, 1962). Texture, and soil organic carbon (C) vary with landscape position. Soil organic C has a regulatory influence on soil N, P, and S availability in the sub-humid tropics (Janssen et al., 1990). Soil nutrient status characterization provides a foundation for understanding how to strategically apply fertilizer and organic inputs, and it is a step towards revised soil management recommendations. The objectives of this paper were to: i) characterize soil nutrient status (P, Zn, K, Ca), and pH in key agricultural regions of Central, Southern and Northern Malawi, ii) quantify the relationship of organic C to other soil characteristics, and iii) develop a spatial understanding of soil characteristics.

MATERIALS AND METHODS

A soil sampling exercise was initiated in 1993 by the Maize Commodity Team in collaboration with the and Plant Analysis Team, Department of Agricultural Research, Malawi Ministry of Agriculture. In cooperation with Extension staff field assistants and district officers, soil samples were collected from over 1,500 smallholder arable fields distributed throughout the country. Information collected at each site included farmer name, village location, cropping system for the last two years, fertilizer use history, and slope category. The majority Downloaded By: [Wageningen UR] At: 14:24 2 March 2011 of fields chosen had not been fertilized in the previous cropping season. Fields sampled were representative of the dominant cropping systems used by smallholders, compared to a farm survey conducted by Jones and Sakala (1991). Evaluation of responses on the soil sample data sheet showed that 80% of smallholder fields sampled were used to grow maize as a sole crop, or maize intercropped with minor. The major sole crops in the remainder of fields sampled were tobacco (Nicotiana tabacum L.), groundnuts (Arachis hypogaea L.), and cotton (Gossypium hirsutum L.). The agroecosystems of Malawi are dominated by a sub-humid tropical, uni- modal rainfall system, where rainfall varies from 700-1,400 mm annually. Land preparation in Malawi is almost entirely by hand-held hoes. In the South, a couple months after the end of the growing season in July or August, residues and weeds are incorporated in the furrow between ridges. New ridges are then build upon the former furrow location, in preparation for the next growing season. Ridges Downloaded By: [Wageningen UR] At: 14:24 2 March 2011

FIGURE 1. Map of Malawi showing the boundaries of the eight agricultural development divisions (ADD). Soil sampling was conducted in five ADDs: Karônga and Mzuzu in the North, Lilongwe and Kasungu in the Center, and Blantyre in the South. The shaded area on the map is Lake Malawi. SOIL NUTRIENT STATUS OF SMALLHOLDER FARMS 2575

are planted with the onset of reliable rains, usually around December. The rainy season lasts about three to four months. In the Center and North of Malawi, a similar land preparation and growing season schedule is followed. However, residues are almost always burnt rather than incorporated as they are in the South. A composite soil sample of 10 sub-samples was collected randomly from the 0- 20 cm depth of the ridge. Furrow soil was not sampled. Ridge soil has been shown to represent the soil environment which crop plants have access to in the Malawi smallholder cropping system (Snapp, unpublished data). Soil was air dried, ground to pass a 2-mm sieve, and analysis conducted. Mehlich 3-extractable P, Zn, K, and Ca were determined, an ammonium fluoride (NH4F) and EDTA based extradant which is highly correlated with extractants widely used in the tropics, Bray 1 and EDTA (Mehlich, 1994; Wendt, 1995). Soil pH was determined in 1:2.5 soil/water ratio. Organic C was determined by a wet digestion and colorimetric method (Anderson and Ingram, 1989). Texture was analyzed by dispersal and hydrometric readings (Anderson and Ingram, 1989). All locations where soil samples were collected were geo-referenced by extension staff using grid overlays on maps. The UTM coordinates systems was used. A database of soil and site characteristics was developed. Over 1,500 fields were sampled, but the geo-referencing system was suspect in some cases (a geopositioning system, GPS, has recently replaced the map grid overlay system for greater reliability). A subset of 1,130 soil samples were identified as reliable from the Mzuzu, Karonga, Kasungu, Lilongwe, and Blantyre administrative units (agricultural development divisions, ADD) in Malawi. The first two districts represent the North, the second two districts the Center, and the final district the Southern region of Malawi (Figure 1). Nutrient status, soil texture, and slope at each site sampled was used to develop descriptive statistics for soil samples from these districts which represents about 60% of the land area of Malawi. Descriptive statistics (mean, mode, and range) from the soil sample database Downloaded By: [Wageningen UR] At: 14:24 2 March 2011 were analyzed based on location, for the five agricultural development divisions, and for 88 natural regions in Malawi. The natural regions were defined by agroecological groups as defined through physiography, climate, and land suitability classes by the Lands Evaluation Project (Venema, 1990). The mean values and standard deviation for soil characteristics for each natural region was evaluated to detect locations which were above or below the mean for the country. A positive, stepwise multiple regression to determine characteristics which predict organic C was carried out for each district and for all five districts together (Statistica, 1993). The database is in the process of being continually updated, and it can be related to socio-economic or other biophysical databases using the site location information (W. Makumba, personal communication, 1997). This soil database has been used to chose representative on-farm sites to locate nutrient response trials (Snapp and Benson, 1995). 2576 SNAPP

TABLE 1. Descriptive statistics for soil and site characteristics of 460 samples from smallholder fields in Mzuzu agricultural development division (Mzuzu ADD) and Karonga agricultural development division (Karonga ADD) in Northern Malawi. Critical values used: , 85%; organic C, 0.8%; pH, 5.2; Zn, 0.5 mg kg"1; Ca, 0.2 -l 1 1 cmolc kg ; K, 0.2 cmolc kg' ; and P, 13.0 mg kg" . Multiple regression of soil characteristics which contribute to organic C status showed that in Mzuzu ADD soil texture was the only significant variable [(R=0.40) Y=12.8+(-0.13*sand % X)] and in Karonga ADD soil texture and Ca were significant variables [(R=0.30) Y=4.9+(-0.03*sand % X)+(0.06*Ca X)].

%> Std. Corr. R Characteristic Critical Mean dev. Median Minimum Maximum org. C

Mzuzu ADD Slope 5.1 4.6 4.9 1.0 19.0 ns Sand % 87.2 75.1 10.4 74.0 47.0 90.0 0.40 Organic C % 72.4 1.2 0.4 1.1 0.2 3.0 na PH 87.1 5.7 0.5 5.7 4.4 7.7 ns Zn mg kg'1 92.5 1.4 1.4 1.0 0.2 5.2 ns 1 Ca cmolc kg" 98.2 2.3 1.1 2.0 0.1 5.5 ns 1 K cmolc kg 74.1 0.5 0.5 0.3 0.1 5.1 ns P mg kg1 57.9 21.7 23.9 14.3 1.0 184.1 ns Karonea ADD Slope 6.3 4.5 5.5 1.0 19.0 ns Sand % 88.9 76.1 10.6 78.0 52.0 94.0 0.15 Organic C % 80.1 1.3 0.5 1.1 0.4 3.3 na pH 87.9 6.2 0.8 6.1 4.0 8.3 ns Zn mg kg1 94.0 1.8 1.2 1.5 0.0 6.4 ns 1 Ca cmolc kg 99.5 3.1 2.2 2.5 0.1 6.2 0.07 K cmolckg"' 88.5 0.7 0.4 0.6 0.1 2.5 ns P mg kg-1 58.8 30.1 29.6 20.2 0.1 163.0 ns Downloaded By: [Wageningen UR] At: 14:24 2 March 2011 •na=not applicable, ns=not significant.

RESULTS AND DISCUSSION Acidity Severe soil acidity is not a major edaphic problem for smallholder fields in Malawi. Moderate soil acidity may constrain yields in some areas. The overall mean pH value was 5.8 (Tables 1, 2, and 3). This supports earlier studies in Malawi which report average soil pH values in the range of 5.7 to 6.0 (Weil and Mughogho, 1993; Wendt, 1993; Young and Brown, 1962). The critical value for acidity below which maize production is affected in the tropics has been shown to be between 5.3 and 5.1 (Landon, 1991). Soil acidity did not reduce maize yields SOIL NUTRIENT STATUS OF SMALLHOLDER FARMS 2577

TABLE 2. Descriptive statistics for soil and site characteristics of 450 samples from smallholder fields in Lilongwe agricultural development division (Lilongwe ADD) and Kasunga agricultural development division (Kasunga ADD) in Central Malawi. Critical soil test values used: sand, 85%; organic C, 0.8%; pH, 5.2; Zn, 0.5 mg kg"1; Ca, 0.2 1 1 1 cmolc kg' ; K, 0.2 cmolc kg" ; and P, 13.0 mg kg' . Multiple regression of soil characteristics which contribute to organic C status showed that in Lilongwe ADD soil texture and Ca were significant [(R=0.43) Y=3.3+(-O.O3*sand % X)] and in Kasunga ADD soil texture, Ca, and slope were significant [(R=0.46) Y=6.6+(-0.07*sand % X)+(0.03 slope % X)+(0.01*CaX)].

%> Std. Corr. R Characteristic Critical Mean dev. Median Minimum Maximum org. C

Lilonewe ADD Slope 4.7 4.5 4.5 1.0 19.0 ns Sand % 84.8 74.7 9.2 78.0 52.0 92.0 0.25 Organic C % 90.6 1.6 0.6 1.5 0.4 3.3 na PH 66.1 5.4 0.4 5.4 4.7 6.8 ns Zn mg kg-1 97.1 1.4 1.0 1.3 0.2 6.0 ns 1 Ca cmolc kg' 100.0 4.1 2.8 3.3 0.8 21.0 0.15 1 K cmolc kg' 88.5 0.4 0.2 0.4 0.1 1.1 ns P mg kg'1 64.3 31.3 29.7 21.4 0.6 130.2 ns Kasunça ADD Slope 4.8 4.4 4.5 1.0 19.0 0.08 Sand % 88.8 78.1 9.1 78.0 58.0 98.0 0.11 Organic C % 91.8 1.7 0.8 1.6 0.3 4.6 na pH 99.2 5.9 0.4 5.9 4.3 7.2 ns Zn mg kg'1 91.9 1.5 1.9 1.2 0.1 5.7 ns 1 Ca cmolc kg" 100.0 1.9 0.8 1.7 0.2 6.0 0.05 1 K cmolc kg' 83.6 0.5 0.4 0.4 0.1 1.5 ns P mg kg'1 59.3 24.2 21.5 17.5 0.1 81.0 ns Downloaded By: [Wageningen UR] At: 14:24 2 March 2011 "na=not applicable, ns=not significant.

in 20 on-farm trials where pH was at least 5.2 (W.D. Sakala, J.D.T. Kumwenda, and Snapp, unpublished data, 1997). On this basis, 5.2 was chosen as the critical value to evaluate pH data in this study. The pH was >5.2 for 87% of the soils tested in the North (Table 1) and for 78% of the soils in Southern Malawi (Table 3). In Lilongwe ADD, which contains the Central mid-altitude plain of Malawi, only 66% of soils tested were >5.2 (Table 2). This is in sharp contrast with Kasungu ADD, also in Central Malawi, where over 90% of pH values were greater than 5.2. The average acidity of soils in Lilongwe ADD was 0.3 units lower than any other region. This was somewhat surprising given that Alfisols are the dominant soil type in Lilongwe ADD. 2578 SNAPP

TABLE 3. Descriptive statistics for soil and site characteristics of 220 samples from smallholder fields in Blantyre agricultural development division (Blantyre ADD) in Southern Malawi. Critical soil test values used: sand, 85%; organic C, 0.8%; pH, 5.2; 1 1 1 1 Zn, 0.5 mg kg ; Ca, 0.2 cmolc kg ; K, 0.2 cmolc kg ; and P, 13.0 mg kg- . Multiple regression of soil characteristics which contribute to organic C status showed that only soil texture was a significant variable [(R=0.43) Y=14.6+(-0.16*sand % X)].

%> Std. Corr.R Characteristic Critical Mean dev. Median Minimum Maximum org. C

Slope 7.4 5.8 4.5 1.0 19.0 ns Sand % 80 72.3 11.5 70.0 52.0 90.0 0.43 Organic C % 77 1.2 0.5 1.1 0.3 2.9 na PH 78 5.9 0.4 5.8 5.1 6.9 ns Zn mg kg-l 90 1.7 1.3 1.3 0.0 9.6 ns 1 Ca cmolc kg' 99 3.0 2.1 2.4 0.2 9.5 ns -l K cmolc kg 99 1.1 0.9 0.9 0.0 13.0 ns P mg kg"1 77 45.5 42.3 38.2 0.9 119.0 ns "na=not applicable, ns=not significant.

Agronomic practices may have contributed to soil acidity. For example, widespread burning of residues, continuous maize production without fallowing land or rotating crops, and use of acidifying fertilizers (Jones and Sakala, 1991). Average pH values were higher in the North than in the rest of Malawi; however, the highest variability of pH and the lowest pH (4.0) also occurred in the North. This is presumably due to localized soils found in the Northern region (Young and Brown, 1962). Oxisols are highly weathered and generally acidic. Downloaded By: [Wageningen UR] At: 14:24 2 March 2011 Excess hydrogen (H+) ions are not necessarily responsible for yield reductions in acid soils. Acidity effects are mediated by a complex of aluminum (Al) toxicity and nutrient imbalances. Aluminum becomes available if pH falls below about 5.4 and causes toxicity problems in soils with acidity <4.8, if exchangeable Al is high. Aluminum toxicity has rarely been observed in Malawi, although it may occur in Southern Malawi tea estates and in isolated high altitude areas in Central Malawi (Brown and Young, 1966; Aggarwal et al., 1997). Moderate acidity is widespread in Malawi, in contrast to the small proportion of very acid sites (Tables 1,2, and 3). Amendment of soil pH with lime has not been shown to consistently enhance yields of maize for moderately acid soils (Landon, 1991; Young and Brown, 1966). Use oflime is in any case prohibitively expensive due to transport costs for the vast majority of smallholder farmers. Yet, it is possible that the wide-spread moderate acidity in Malawi is associated with reduced N and P availability. Janssen et al. (1990), working in Kenya on the QUEFTS model, found support for a negative relationship of pH to N and P uptake SOIL NUTRIENT STATUS OF SMALLHOLDER FARMS 2579

by maize at moderate acidity levels. However, a global soil and plant nutrition survey found little relationship of pH to N and P availability (Sillanpaa, 1982). The extent to which moderate acidity influences performance of major crops (e.g., maize, groundnut, and tobacco) in Malawi deserves more research attention, with a focus on organic inputs to ameliorate soil pH and nutrient availability. Lilongwe ADD includes the fertile Central plain which is responsible for about 40% of cereal grain production in Malawi. The diagnosis of moderate to severe acidity in this "bread basket" of Malawi is a cause for concern (Table 2). Organic technologies such as mulching and incorporation of organic materials show potential as alternatives to liming, to ameliorate soil pH, and enhance nutrient availability in acidic soils (Pocknee and Sumner, 1997; Ruhigwa et al., 1993). Mulch added at 21 ha1 to an Oxisol soil in Nigeria increased pH by 0.1 unit, and available P by 40%. Mulch added at 61 ha1 was even more effective, it increased pH by 0.5 units and available P by 85% (Lai, 1981). Cation and N content of organic inputs has been shown to be closely related to ability to reduce soil acidity and enhance soil N availability (Pocknee and Sumner, 1997). This relationship remains to be investigated for soils and residues of smallholder farms in Malawi.

Texture Sand content in the Center and Northern regions was about 76% (Tables 1 and 2). Average sand content in the Southern region was slightly lower, 72% (Table 3). These results concur with descriptions of Malawi soils as generally sandy in texture and well drained (Brown and Young, 1966; Wendt, 1993; Young and Brown, 1962). Soil samples were obtained from the primary maize production fields of smallholder agriculture. Average sand content reported here would have been much lower if the hydromorphic soils along drainage lines had been included in the sample. These soils are used for grazing livestock and winter vegetable

Downloaded By: [Wageningen UR] At: 14:24 2 March 2011 production, and they have a high content in the range of 30 to 60% (Young and Brown, 1962; Brown and Young, 1966). Soil texture class was determined for the soils reported here and soils in the North were largely sandy loams or sands. The remainder were sandy loams, and sandy clay loams. A similar pattern was observed in the Center and South, although sandy loams were generally much more common than in the North. Soils which are more than 85% sand tend to be excessively drained, low in soil organic C, and very low in nutrient content (Landon, 1991). Tables 1, 2, and 3 show that about one-sixth of the soils sampled had sand content >85%. Water and N losses can be high form these soil types. Results from a simulation model and field trials conducted in Central Malawi suggest that N losses from in sandy soils are on the order of 50 kgN ha1 annually, compared to approximately 20 kg N ha1 annually from sandy soils (Thornton et al., 1995). This indicates the importance of agronomic technologies to reduce N and water loss. Application of N fertilizer using multiple splits, adjusted for the rainfall pattern, is a promising approach shown to reduce N losses in nearby Zimbabwe (Piha, 1993). 2580 SNAPP

Organic Carbon The Northern and Southern regions of Malawi had the lowest soil organic C, on average 1.2% (Tables 1 and 3). Soil organic C values of the Center of the country were higher, about 1.6% (Table 2). Weil and Mughogho (1993) and Wendt ( 1995) also found mean organic C levels of about 1.5% in soils from Malawi. Minimum and maximum values across the country ranged from 0.2 to 4.6% organic C. This raises the question of what is an adequate level of soil organic C. Data from long- term trials in the sub-humid tropics suggest that a level of soil organic C between 0.8% to 1% is sufficient to support crop production and prevent soil degradation (Araki, 1993; Pieri, 1995). This is for soils with sand content above 60%, as is typical of the topsoil in Malawi. This is a minimal soil organic C level, below which soil structure is severely impaired. Based on this rational the critical value of 0.8% was used to evaluate the soil organic C data. About 80% of the soils tested had organic C values above the critical value of 0.8% (Tables 1,2, and 3). The mean value for natural regions with low organic C levels was 1.0%, and thus just above 0.8% (Table 4). The data indicate that the majority of smallholder fields in Malawi have adequate organic C to prevent soil physical degradation. This suggests that agronomic technologies to enhance soil organic C levels are not recommended on a large scale. Although, erosion prevention measures must be taken to preserve the remaining soil and the wide-spread practice of residue burning evaluated for its long-term effects on soil C. As a cautionary note, the levels observed here on smallholder farms were severely depleted compared to the 1.8 to 3.5% organic C found under natural vegetation in Malawi and at research stations such as Chitedze (Brown and Young, 1966; Snapp et al., 1998). Building has been advocated as a critical need throughout sub-Saharan Africa. However, it is difficult to produce sufficient organic materials

Downloaded By: [Wageningen UR] At: 14:24 2 March 2011 to increase soil organic matter in the semi-arid and sub-humid tropics, given that land and labor are often in short supply and the rapid C mineralization rate in this environment (Snapp et al., 1998). A more realistic goal may be to promote incorporation of selected legumes in smallholder cropping systems. This will increase the amount of high quality organic residues and N from biological nitrogen fixation and enhance crop productivity and biological nutrient cycling efficiency, without necessarily increasing soil organic C. Notable exceptions to this strategy were hilly regions in Ntcheu, Nkhotakota and Michinji. These were natural regions with very low C (Table 4), where management to improve soil organic C levels should be implemented. Multiple regression analysis was used to evaluate the relationship of organic C to other soil characteristics for the entire soil database. Positive, stepwise multiple regression showed that sand and calcium were significant contributors towards predicting organic C, where R=0.61 [Y=4.4+(-0.04*XS)+(0.03*XCa) and 1 Y=organic C (%), where: XS=sand (%), and XCa=Ca (cmolc kg" )]. Sand, and to ai O

TABLE 4. Soil characteristics and distinctive groupings higher and lower than countryside means by natural regions in Malawi. United States Department of Agriculture terminology equivalent for FAO system soil family names is Udic Rhodustalfs and Oxic Haplustalfs for Chromic Luvisol and Haplic Lixisols, respectively. 1 Soil analyses Moisture Elevation Soil class O Natural region description (mean) (mm) (m) (FAO system)

Downloaded By: [Wageningen UR] At: 14:24 2 March 2011 > Lilongwe Plain High org. C 2.0% 850 1,200 Chromic Luvisol Upper Bua Plains High org. C 2.5% 900 1,150 Chromic Luvisol, Haplic Lixisol Michinji Hills Low org. C 1.1% 800 1,310 Haplic Lixisol, Gleyic o a Nkhotakota Scarp Zone Low org. C 1.1% 1,130 1,230 Haplic Lixisol, Chromic Luvisol m Ntcheu Scarp Zone Low org. C 0.9% 1,020 1,230 Haplic Lixisol, Chromic Luvisol 73 Ntcheu Region Low org. C 1.0% 940 970 Haplic Lixisol, Chromic Luvisol Chikwawa-Thyolo Escarpment Low P 15 mg kg1 960 480 Eutric Cambisol, Chromic Luvisol Chileka Region Low P 5 mg kg"1 980 690 Chromic Luvisol, Eutric Cambisol Michinji Hills Low Zn 1.1% 800 1,310 Haplic Lixisol, Gleyic Cambisol Dedza Hills Low Zn 0.9 mg kg1 910 1,340 Haplic Lixisol, Chromic Luvisol Dedza Scarp Zone Low Zn 0.8 mg kg'1 900 860 Haplic Lixisol, Chromic Luvisol

to 00 2582 SNAPP

some extent Ca, also were predictors of soil organic C when the regression analysis was conducted separately by agricultural district. There was one addition, a slope variable from the site description characteristics, which was a significant variable predicting soil organic C in the Kasungu district (Table 2). Overall the influence of Ca was limited; sand content was the major predictor of soil organic C content. Sand is negatively related to organic C (Figure 2). Sandy soils are generally low in soil organic C due to the lack of physical and chemical protection from mineralization processes (Parton et al., 1987; Pieri, 1995). Soil Ca availability has not been generally found to be related to organic C. However, organic C is related to soil cation exchange capacity (Landon, 1991). This may be a factor influencing the relationship of Ca and organic C observed here.

Phosphorus Soil P status in Malawi was generally high, about 60% of soils were above the critical value of 15 mg kg'1 (Tables 1, 2, and 3). This finding is supported by a global survey of soils which indicated that, in general, Malawi soils are sufficient in P for maize production, and have a higher P status than surrounding countries (Sillanpaa, 1982). There were two natural regions in the South which were markedly low in P, Chileka and Chikwawa-Thyolo (Table 4). The variability is demonstrated by a comparison of means and standard deviations for P in the North (Table 1 ), Center (Table 2), and the South (Table 3). The standard deviation was as high as the mean for most of the country. Variability of extractable P in Malawi has been noted by others, such as Young and Brown (1962) and Wendt (1993). This variability is not surprising due to the tremendous spatial heterogeneity of topography in Malawi. R. Weil and S. Mughogho (USAID final report, 1992) reported extensive inter-fingering of light

Downloaded By: [Wageningen UR] At: 14:24 2 March 2011 and dark red soils at a scale of less than 0.5 kilometers. Further, average smallholder field sizes are in the range of 0.5 to 1.5 hectares in Central Malawi, and smaller in South (Snapp et al., 1998). This leads to a wide range of fertilizer and agronomic management histories across the landscape. It is also must be considered that the extractable P values reported here were not correlated closely with crop response to P fertilizer. Less than one-third of on-farm trials showed the expected response of maize to P fertilizer for soils with low extractable P (Chilimba, 1996). Wendt (1993) found similar results to Chilimba: available soil P, as indicated by extractants Mehlich 3, Bray 1, and Olsen, identified correctly the majority of P sufficient sites, but not low P sites. Only a few fields showed maize responses to P fertilizer, and 20 kg ha1 phosphate was the maximum level at which a maize response was obtained. The data presented here and the literature on maize response to P fertilizer both suggest that P response and soil P levels are highly variable. However, crop response to P and available soil P do not always concur. SOIL NUTRIENT STATUS OF SMALLHOLDER FARMS 2583

Regression 65 75 95% confid. Sand (%)

FIGURE 2. A linear regression of soil samples from Central Malawi showing the negative relationship of sand content soil and organic C.

The generally sufficient P status in Malawi supports recent efforts to evaluate 1 blanket fertilizer recommendations for hybrid maize of 45 kg P2O5 ha , and develop lower P recommendation rates (Kumwenda et al., 1997). One problem with over application of P has been noted by Wendt (1993, 1995): P fertilizer can reduce Zn availability under Zn-deficient conditions. This potentially is a serious problem in low Zn areas such as the Dedza region in Central Malawi (Table 4). Phosphate fertilizer will generally still be required for high P demanding crops such as common bean. This was shown by the application of small amounts of P, e.g., 20

Downloaded By: [Wageningen UR] At: 14:24 2 March 2011 1 kg P205 ha" , to maize/bean intercrops which consistently enhances yields by 30 to 100% in Central Malawi (Aggarwal et al., 1997).

Potassium Potassium levels in the North and Center were over the critical value for about 80% of the soils tested (Tables 1 and 2). Potassium status was even higher in the South where soil analysis indicated that over 99% of soils had sufficient K (Table 3). The mean K value for the soils tested here was about 3 fold higher than the critical value. Generally, maize fertilizer trials have indicated no response to K in Malawi smallholder fields (Kumwenda et al., 1997). The data presented here and the literature indicate that Malawi soils appear to be broadly K sufficient.

Calcium Almost 100% of soils in Malawi have sufficient Ca as indicated by soils analysis (Tables 1, 2, and 3). Studies have indicated low cation exchange capacity in 2584 SNAPP

Malawi soils, a characteristic which was not studied here, but one expected to influence Ca availability (Mwandemere and Robertson, 1975). Further, the Ca critical value used here was developed for maize. Crops such as groundnuts have a high demand for Ca and some soils in Malawi may be Ca deficient with regards to groundnut production. Calcium supplying soil amendments, gypsum and lime, have been shown to be inconsistent in enhancing groundnut pod filling and yields in Central Malawi (Nyirenda et al., 1992). Calcium uptake in groundnuts is influenced by water availability as well as soil Ca supply. This may explain the uneven response of groundnuts to Ca nutrition in Malawi. Overall, soil Ca status was very high in the soils studied here and amendment with Ca does not appear to be required.

Zinc Zinc appears to be sufficient throughout most of Malawi. Yet, soil Zn status was highly heterogeneous, as indicated by high standard deviations. Mean values for Zn were above the critical level of 0.8 mg kg'1 for about 94% of soils in the North and Center (Tables 1 and 2), and 90% in the South (Table 3). A global study by Sillanpaa ( 1982) indicated mean Zn values which were generally sufficient in Malawi, however, the study also indicated isolated areas of very Zn deficient. In Malawi, Zn fertilizer responses in maize fertilizer trials have been noted for the Dedza hills region (Kumwenda et al., 1997; Wendt, 1993). The same region was identified as Zn deficient through statistical analysis of soil Zn per natural regions (Table 4). Soil Zn values for Dedza region were about 0.8 mg kg"1, whereas values for the rest of the country were almost double this. The soil parent material in Dedza is igneous granite, generally low in total Zn (less than 40 mg kg1), which could help explain low availability of Zn in sandy soils of the region (Tisdale et al., 1985). Another important factor could be the cool

Downloaded By: [Wageningen UR] At: 14:24 2 March 2011 temperatures in this high altitude region. Cool temperatures tend to slow plant growth and enhance Zn deficiency (Tisdale et al., 1985). Taken together, the data confirmed earlier suggestions that Zn fertilizer should be made available to smallholder farmers in the Dedza area (Wendt, 1993).

CONCLUSIONS

The majority of Malawi soils were loamy sands and moderately acid, with sufficient to low nutrient levels. Organic C levels were adequate to maintain soil structure in the main, although much reduced from the non-cultivated state. Low soil organic C suggested that N supply was inadequate throughout most of the country. Inadequate N supply was also indicated by the literature which demonstrated consistent crop responses to N (Brown and Young, 1966; Chilimba, 1996). Widely deficient N status was in contrast to soil K and Ca levels which were sufficient throughout most of the country. Soil P, on the other hand, was highly SOIL NUTRIENT STATUS OF SMALLHOLDER FARMS 2585

variable. The mean P status in Malawi was well above soil P levels generally observed for soils in Southern Africa. This indicates P-sufficient soils, an observation supported by the limited response of maize to P fertilizers in Malawi (Wendt, 1993). However, the data reported here indicated that local P deficiencies were severe in some areas of Malawi. Soil Zn was also heterogeneous. In general, soil Zn was adequate, but there were isolated areas of severe Zn deficiencies. Zinc deficiencies occurred in cooler, high altitude areas with igneous granite parent material. Soil characteristics reported here were typical of the range reported previously for soils in Malawi. The data highlighted the need for action on fertilizer recommendations: i) Zn fertilizers should be made available to farmers in the Zn deficient Dedza region, ii) generally observed high soil P status strengthened earlier reports by Wendt (1993) and Kumwenda et al. (1997) that new fertilizer recommendations should halve P fertilizer rates (e.g., a reduction from 45 to 23 kg phosphate per ha for hybrid maize), and iii) the extent of low organic C, sandy soils reinforced the importance of fertilizer recommendations which improve N efficiency. This includes new, targeted, and flexible approaches to fertilizer use and developing alternative, legume-based N sources for smallholder farmer. Quality of organic matter and nutrient cycling efficiency are areas of active investigation in Malawi (Jones et al., 1996; Snapp et al., 1998). More attention is warranted to use of combined inorganic and organic fertilizers, and split applications of small amounts of fertilizer. New findings reported here were the broad extent of moderate soil acidity in the Lilongwe plain and the delineation of natural regions with low C and low P. The data suggest that new research initiatives should be undertaken on the use of organic materials to ameliorate moderate acidity and enhance N supply in Central Malawi. Technologies are required to address problem soils with very low organic C and P, including combined use of inorganic and organic fertilizers. Priority Downloaded By: [Wageningen UR] At: 14:24 2 March 2011 should be given to organic materials which can be easily produced by smallholder farmers and by-products of crops widely grown, such as residues from groundnuts.

REFERENCES

Aggarwal, V.D., S.K. Mughogho, R.M. Chirwa, and S.S. Snapp. 1997. Field-based screening methodology to improve tolerance of common bean to low-P soils. Commun. Soil Sci. Plant Anal. 28(17&18):1623-1632.

Anderson, J.M. and J.S.I. Ingram. 1989. Tropical and Fertility: A Handbook of Methods. CAB International, Aberystwyth, UK.

Araki, S. 1993. The effect of burning, ash fertilization and soil organic matter on productivity of the Chitemene shifting cultivation in Northern Zambia, pp. 367-375. In: K. Mulongoy and R. Merckz (Editors), Soil Organic Matter Dynamics and Sustainability of Tropical Agriculture. John Wiley and Sons, Chichester, England. 2586 SNAPP

Brown, P. and A. Young. 1966. The physical environment of Central Malawi: With Special Reference to Soils and Agriculture. Government Printer, Zomba, Malawi.

Chilimba, A.D.C. 1996. The effectiveness of universal soil extractants in detecting nutrient deficiencies in Malawi soils. M.Sc. thesis, Crop Science Department, Bunda College of Agriculture, University of Malawi, Lilongwe, Malawi.

Heisy, P.W. and W. Mwangi. 1996. Fertilizer use and maize productivity, pp. 193-211. In: D. Byerlee and C.K. Eicher (eds.), Africa's Emerging Maize Revolution. Lynne Publishers, Boulder, CO.

Janssen, B.H. 1991. Integrated nutrient management: the use of organic and mineral fertilizers, pp. 89-105. In: H. van Reuler and W.H. Prins (eds.), The Role of Plant Nutrients for Sustainable Food Crop Production in Sub-Saharan Africa. Dutch Association of Fertilizer Producers (VKP), Liedschendam, The Netherlands.

Janssen, B.H., F.C.T. Guiking, D. van der Eijk, E.M.A. Smaling, J. Wolf, and H. van Reuler. 1990. A system for quantitative evaluation of the fertility of tropical soils (QUEFTS). Geoderma 46:299-318.

Jones, R.B. and W.D. Sakala. 1991. An informal survey of maize agronomic practices in Kasungu and Lilongwe agricultural development divisions, Malawi. Fanning Systems Research and Extension Newsletter, CIMMYT 6:11-17.

Jones, R.B., S.S. Snapp, and H.S.K. Phombeya. 1997. Management of Leguminous leaf residues to improve nutrient efficiency in the sub-humid tropics, pp. 239-252. In: K. Giller and G. Cadisch (eds.), Driven by Nature: Plant Litter Quality and Decomposition. CAB International, Wallingford, UK.

Kumwenda, J.D.T., S.R. Waddington, S.S. Snapp, R.B. Jones, and M.J. Blackie. 1997. Soil fertility management in Southern Africa, pp. 153-172. In: D. Byerlee and C.K. Downloaded By: [Wageningen UR] At: 14:24 2 March 2011 Eicher (eds.), Africa's Emerging Maize Revolution. Lynne Publishers, Boulder, CO.

Lal, R. 1981. Management of soils for continuous production: controlling erosion and maintaining physical condition, pp. 188-201. In: D.J. Greenland (ed.), Characterization of Soils: Examples from Some Areas of the Humid Tropics. Clarendon Press, Oxford, UK.

Landon, J.R. 1991. Booker Tropical Soil Manual. Longman Science and Technical, Essex, England.

Mandemere, H.R. and W.K. Robertson. 1975. Selected mineralogical and chemical properties often soils of Malawi. Soil Crop Sci. Soc. Florida Proc. 35:155-160.

Mehlich, M. 1984. Mehlich 3 soil test extractant: A modification of the Mehlich 2 extractant. Commun. Soil Sci. Plant Anal. 15:1409-1416. SOIL NUTRIENT STATUS OF SMALLHOLDER FARMS 2587

Nyirenda, N.E., T.J. Cusack, and V.W. Saka. 1992. Groundnut agronomy research in Malawi: Past achievements and future priorities. In: R.C. Nageswara Rao and P. Subrahmanyam (eds.), Proceedings of the 5th Regional Groundnut Workshop for Southern Africa. International Crops Research Institute for the Semi-Arid Tropics, Patancheru, India.

Parton, W.J., D.S. Schimel, C.V. Cole, and D.S. Ojima. 1987. Analysis of factors controlling soil organic matter in Great Plains grasslands. Soil Sci. Soc. Am. J. 51:1173- 1179.

Pieri, C. 1995. Long-term soil management experiments in semi-arid Francophone Africa, pp. 225-266. In: R. Lal and B.A. Stewart (eds.), Soil Management Experimental Basis for Sustainability and Environmental Quality. Advances in Soil Science, CRC Lewis Publishers, Boca Raton, FL.

Piha, M.I. 1993. Optimizing fertilizer use and practical rainfall capture in a semi-arid environment with variable rainfall. Exp. Agric. 29:405-415.

Pocknee, S. and M.E. Sumner. 1997. Cation and nitrogen contents of organic matter determine its soil liming potential. Soil Sci. Soc. Am. J. 61:86-92.

Ruhigwa, B.A., M.P. Gichuru, N.M. Tariah, N.O. Isirimah, and D.C. Douglas. 1993. Spatial variability in soil chemical properties under hedgerow intercrop species on an acid . Exp. Agric. 29:365-372.

Sillanpaa, M. 1982. Micronutrients and the Nutrient Status of Soils: A Global Study. FAO Soils Bull. 48. Food and Agriculture Organization of the United Nations, Rome, Italy.

Snapp, S.S. and T. Benson. 1995. GIS to develop area-specific fertilizer recommendations in Malawi: Soil resources, yield potential and decision trees, p. 301. In: Agronomy Downloaded By: [Wageningen UR] At: 14:24 2 March 2011 Abstracts. American Society of Agronomy, Madison, WI.

Snapp, S.S., P.L. Mafongoya, and S. Waddington. 1998. Organic matter technologies to improve nutrient cycling in smallholder cropping systems of Southern Africa. Agric. Ecosys. Environ. (In press).

Statistica. 1993. Statistica for Windows Program. StatSoft, Inc., Tulsa, OK.

Thornton, P.K., A.R. Saka, U. Singh, J.D.T. Kumwenda, J.E. Brink, and J.B. Dent. 1995. Application of a maize crop simulation model in the Central region of Malawi. Exp. Agric. 31:213-226.

Tisdale, S.L., W.L. Nelson, and J.D. Beaton. 1985. Soil Fertility and Fertilizers. 4th ed. Macmillan Publishing Co., New York, NY. 2588 SNAPP

Venema, J.H. 1990. Land resource evaluation project. Methods of description, classification and mapping of natural regions of Malawi. Field document No. 3. Malawi Government Ministry of Agriculture Land Husbandry Branch, United Nations Development Programme, Food and Agriculture Organization of the United Nations, Rome, Italy.

Weil, R.R. and S.K. Mughogho. 1993. Nutrient cycling by Acacia albida (Syn. Faidherbia albida) in agroforestry systems, pp. 97-108. In: Technologies for Sustainable Agriculture in the Tropics. ASA Spec. Publ. No. 56. American Society of Agronomy, Madison, WI.

Wendt, J.W. 1993. Diagnosis of regional topsoil nutrient deficiencies in Malawi. Proceedings of the Southern Africa Farming Systems-Extension Conference, 1-3 June 1993, Ezulwini, Swaziland. CIMMYT, Harare, Zimbabwe.

Wendt, J.W. 1995. Evaluation of the Mehlich 3 soil extractant for upland Malawi soils. Commun. Soil Sci. Plant Anal. 26:687-702.

Young, A. and P. Brown. 1962. The Physical Environment of Northern Nyasaland: With Special Reference to Soils and Agriculture. Government Printer, Zomba, Malawi. Downloaded By: [Wageningen UR] At: 14:24 2 March 2011