The Influence of Selected Soil Conservation Practices on Soil Properties and Crop Yields in the Usambara Mountains , Tanzania

The Usambara Mountains in Tanzania are severely affected by soil erosion which has led to deterioration of soil properties and reduced crop productivity. Indigenous soil erosion control measures such as miraba which are widely practised in the area have yielded little success. Field plot experiments were laid down in Majulai and Migambo villages from 2011 – 2014 on typical soils of the Original Research Article Mwango et al.; AIR, 3(6): 558-570, 2015; Article no.AIR.2015.051 559 area (Acrisols). The aim was to single out soil properties developed under the studied soil conservation practices and their impact on crop productivity with reference to maize (Zea mays) and beans (Phaseolus vulgaris). Results showed that total N, OC, available P, Ca 2+ , Mg 2+ , K + and Ph were powerful (P = .05) attributes that discriminated conservation measures. Magnitudes of the discriminating attributes followed the trend: miraba with Tughutu (Vernonia myriantha) mulching >miraba with Tithonia (Tithonia diversifolia) mulching > miraba sole > cropl and with no ‘Soil and Water Conservation’ (SWC) measures (control).

area (Acrisols).The aim was to single out soil properties developed under the studied soil conservation practices and their impact on crop productivity with reference to maize (Zea mays) and beans (Phaseolus vulgaris).Results showed that total N, OC, available P, Ca 2+ , Mg 2+ , K + and Ph were powerful (P = .05)attributes that discriminated conservation measures.Magnitudes of the discriminating attributes followed the trend: miraba with Tughutu (Vernonia myriantha) mulching >miraba with Tithonia (Tithonia diversifolia) mulching > miraba sole > cropl and with no 'Soil and Water Conservation' (SWC) measures (control).Contents ofmicro-nutrients did not differ significantly with SWC measures except for Zn which was significantly (P = .05)lowin the control.Bulk density and available moisture content (AMC) were also strong discriminators of conservation measures.Maize and bean grain yields differed significantly (P = .05)withthe trend: miraba with Tughutu > miraba with Tithonia > miraba sole > control in both villages.Crop yields under miraba were a function of AMC and pH (R 2

INTRODUCTION
The problem of soil erosion is global, and has been reported all over the world to affect agricultural sustainability [1,2,3].For example the Usambara Mountains of Tanzania which are characterized by a high population density of about 120.4 persons/km 2 , and practise farming on steep slopes of more than 40% due to land scarcity, suffer from severe soil degradation by water erosion [4,5].Soil loss, nutrient depletion and reduced capacity of the soil to retain water are major forms of soil degradation in the area.These have led to deterioration of soil properties and reduced crop productivity [6].Population pressure in the area has led to increased land use intensity and expansion of cultivation of food and cash crops in valleys and sloping land [4,5].
There is a growing concern that land use practices in the Usambara Mountains may not be sustainable because of their detrimental effects on soil properties [4,7].To address the problem of soil degradation by water erosion, Usambara farmers developed indigenous 'Soil and Water Conservation' (SWC) measures such miraba (rectangular grass bound strips that do not necessarily follow contour lines), micro-ridges and stone bunds as integral part of their farming systems, while introduced measures have often been rejected or minimally adopted because they were expensive in terms of money and labour [9,8].Surprisingly however, the indigenous soil erosion control measures implemented in the area have remained poorly documented [8].Besides, farmers' efforts to conserve the degrading land have yielded very little success, and deterioration of some soil properties are active even in places where SWC measures are practised [9,4,7,10].This is partly due to limited knowledge on the effectiveness of the indigenous SWC practices.Moreover, indigenous SWC measures in the area have been for decades left traditional with little scientific intervention for improvement [9,10].Indigenous SWC measures have been documented to play a considerable role in controlling soil erosion and improving crop yield.For example, stone bunds in Ethiopia have been reported by Van campenhout et al. [11] to be effective in increasing yields from 632 to 683 kg ha -1 for cereals, from 501 to 556 kg ha -1 for Eragrostis tef and from 335 to 351 kg ha -1 for Cicer arietinum as compared to the situation without stone bunds.Likewise the study by Msita [10] in Usambara Moutains, Tanzania revealed miraba to have some contribution in controlling soil erosion and increased maize yield form 0.7 Mg ha -1 in cropland with no soil conservation measures to 1.1 Mg ha -1 in farms with miraba.
Although studies on the effectiveness of some introduced SWC technologies on soil erosion control and agricultural productivity have recently been carried out in Western Usambara Mountains [9,7], the contribution of indigenous SWC measures including miraba which is the most preferred in the study area have not fully been investigated [4,10].Even when investigated, not a single study has attempted to explain the linkages that exist between soil properties and crop productivity associated with SWC technologies.Furthermore, land use planners, agricultural managers and extension officers need sound information to guide implementation of SWC practices within the context of improved soil properties and maximized crop production; yet, at present such information does not exist.
The study reported here in, was therefore aimed at establishing the linkages between identified soil properties associated with soil conservation practices namely miraba and miraba with various mulching materials with reference to productivity of maize (Zea maize) and beans (Phaseolus vulgaris) under smallholder farming conditions in Usambara Mountains.The objectives of this study were (i) to identify soil properties that discriminate between selected SWC practices (ii) to test whether the identified soil properties correlated with crop yield and (iii) to investigate the relation between the identified soil properties and crop yield.

Description of the Study Sites
Migambo and Majulai villages in Western Usambara Mountains, Lushoto District, Tanzania (Fig. 1) are located between 38º15' to 38º24' E and 4º34' to 4º48' S. Migambo is humid cold with mean annual air temperature of 12ºC−17ºC and an annual precipitation ranging from 800-2300 mm.Majulai is dry warm with mean annual air temperature between 16ºC and 21ºC and annual precipitation of 500-1700 mm.The annual evapo-transpiration (ETo) as estimated by the local climate estimator software (New_LocClim) [12] ranges from 100 mm to 145 mm.The Usambara Mountains support a large population density of more than 120.4 persons/km 2 [5].According to the World Reference Base (WRB) for Soil Resources [13] the soils in Majulai site classified as Chromic Acrisols (Humic, Profondic, Clayic, Cutanic, Colluvic) whereas in Migambo site the soils are Haplic Acrisols (Humic, Profondic, Clayic, Colluvic).
The main land uses include cultivation on slopes and valley bottoms, settlements on depressions, lower ridge summits and slopes and forest reserves on ridge summits and upper slopes.Vegetables such as carrots, onions, tomatoes, cabbages and peas are grown as sole crops in valleys under rain fed or traditional irrigation.Beans are grown mainly during long rains and maize in short rains.Irish potatoes and fruits namely peaches, plums, pears, avocado, and banana are grown on ridge slopes under rain fed mixed farming.Irish potatoes are also grown in valleys as sole or intercropped with maize.

Establishment of Miraba in Field Plots
Miraba were established using Napier grass (Pennisetum purpureum) barriers in field plots in April 2011 about nine months before crops were grown.Tillers of Napier grass were planted in single rows at 10 cm spacing perpendicular to the general slope and were maintained to about 50 cm wide strips.Napier grass barriers across the slope were spaced 5 m apart to mimic the recommended maximum effective width of hand made bench terraces [14].On the other hand, the spacing of Napier grass barriers forming miraba along the slope was set at 3 m apart.
It has been documented that soil conservation measures such as Fanya Juu (hillside ditches made by throwing excavated soil on the upslope of the ditch, built along contour lines at appropriate intervals depending on slope gradient) and stone bunds tend to progressively form bench terraces when at close spacing [14,15].Moreover, the closer the grass strips are the more effective they become in controlling soil erosion [15].Progressive bench terrace formation is also possible under miraba when adjusted to appropriate spacing of grass strips.Natural bench terrace formation as a result of miraba implementation is much less expensive compared to mechanical bench terrace construction which is feared by farmers.Bench terraces are highly recommended for use in Usambara Mountains [16,17,9,4].

Experimental Design
Miraba plots 22m x 3 m in a randomized complete block design (RCBD) were set in the lower ridge slopes at 50% slope in Majulai and 45% slope in Migambo village (Fig. 2).Maize and beans were planted in rotation as test crops in 2012 and 2013/14 rain seasons, where maize was planted during short rains (vuli) and beans during long rains (masika).The treatments included plots with (i) Miraba and planted with maize or beans (MI) (ii) Miraba with Tithonia mulching and planted with maize or beans (MITH) (iii) Miraba with Tughutu mulching and planted with maize or beans (MITG) (iv) No SWC measures (CO) (Control) and planted with maize or beans, all replicated three times.

Mulching Materials
Mulching materials used included the leaves of Tithonia diversifolia (Alizeti Pori) and Vernonia myriantha (Tughutu) in both villages.The mulch was applied under miraba two weeks after crops germinated at the rate of 3.6 Mg ha -1 dry weight.These shrubs were chosen as mulches because the plants are readily available in the area and have been documented to contain appreciable amounts of N, P and K [18,6].Samples from each mulching material were collected for determination of total N, available P, K + , Mg 2+ , Ca 2+ and Na + .

Determination of Soil Chemical and Physical Properties
The impact of SWC measures on soil chemical and physical properties was determined by taking composite topsoil samples (0 -30 cm depth) from each treatment for the analysis of pH, OC, total N, available P, K  Undisturbed core soil samples were also collected from 0 -5 cm depth for bulk density and available moisture content determination.Soil samples were collected after every cropping season i.e. long rains and short rains from 2012 to 2013/14.In each runoff experimental site a representative soil profile was excavated and described, and soil samples collected from each horizon for pedological characterisation.Undisturbed core soil samples were taken from 0-5 cm, 45-50 cm and 95-100 cm soil depths by Kopecky's core rings (100 cm 3 ) for bulk density and available moisture determination for further characterization of the representative soil profiles.The soil profiles were classified to tier-2 according to WRB for Soil Resources [13].

Crop Yield Determination
Maize (Zea mays) variety PANNAR 67 and beans (Phaseolus vulgaris) Kilombero variety were planted in runoff plots during the 2012 and 2013/14 rainy seasons with maize in short rains (vuli) and beans during the long rains (masika).The spacing was 75 cm× 30 cm for maize and 50 cm× 25 cm for beans.Beans were always planted three weeks before maize was harvested in Migambo and two weeks in Majulai village.Farmyard manure with 0.6% N, 0.4% P, 0.5% K and 15% OC was basal and spot applied at the rate of 3.6 Mg ha -1 air-dry weight, DAP 18: 46: 0 NPK ratio and Urea 46% N were applied at the rate of 80 kg ha -1 , but Urea was not applied for beans.At maturity maize and bean grains were harvested and dried to about 13% moisture content.

Soil and Plant Samples Analysis
Soil analysis was done following Moberg's Laboratory Manual [19].Organic carbon (OC) was measured using the dichromate oxidation method, total nitrogen (TN) by Kjeldahl method, available phosphorus (Bray-I), exchangeable bases (Ca 2+ and Mg

2+
) by atomic absorption spectrophotometer, exchangeable Na + and K + by flame photometer and pH water by normal laboratory pH meter.The available Fe, Mn, Zn and Cu were extracted using buffered DTPA (Diethylenetriaminepentaacetic acid) method and the DTPA extract was analysed in an Atomic Absorption Spectrophotometer (AAS).Soil texture was determined by Hydrometer method.Bulk density was determined by oven drying and weighing method.Soil moisture retention characteristics were studied using sand kaolin box for low suction values and pressure plate apparatus for higher suction values [20].

Statistical Analysis
Bartlett's test for homogeneity of variance was conducted to test data normality using Gen Stat software [21].All data were subjected to Analysis of Variance (ANOVA).GenStat statistical analysis software [21] was used for the analysis and significant differences were tested by the Least Significant Difference (LSD 0.05 ).Correlation and multiple linear regressions were performed using Minitab software [22] to determine the relationship between soil properties and crop yield under the studied SWC measures.

Selected Chemical Properties of Mulching Materials
Chemical properties of mulching materials are presented in (Table 1).It can clearly be seen that Tughutu had higher nutrient contents than Tithonia (Table 1).This situation is also supported by other researchers [18] who also found higher NPK contents in Tughutu than in Tithonia shrub.3.6 0.3 6.3 1.4 0.9 0.04 Farm yard manure 1.7 0.4 1.9 0.9 0.6 0.07

The Influence of SWC Measures on Selected Soil Physico-chemical Properties
Variability of soil chemical and physical properties between SWC measures are presented in Tables 2 & 3. Considering the soil chemical properties in relation to the SWC measures, most of the properties were significantly (P = .05)different between treatments.The differences can be explained by the influences of the SWC measures applied.It was revealed in both villages that the contents of all studied macro nutrients followed the trend that: miraba with Tughutu mulching > miraba with Tithonia mulching > miraba sole >cropl and with no SWC measures (Table 2) except for Na + which did not significantly (P = .05)differ.Similarly pH followed the same trend.It was therefore concluded that total N, OC, P, Ca 2+ , Mg 2+ , K + and pH were powerful attributes that differentiated SWC measures.Studies by Tenge and Kyaruzi [9, 7] revealed similar observations where terracing such as bench and Fanya Juu terraces effectively control runoff and soil losses, thus improving soil physical and chemical properties in Usambara Mountains.The higher pH and macro nutrient status under miraba with Tughutu mulching than under miraba with Tithonia mulching can be explained by the higher nutrient contents of Tughutu as compared with Tithonia mulching material (Table 1).The higher NPK contents in Tughutu than in Tithonia shrub was also reported by Wickama and Mowo [18].It is also well known that exchangeable bases have strong positive correlation with soil pH [23,24].In the case of micro nutrients, it was found that there were no significant (P = .05)differences between SWC measures except for Zn which was significantly low under cropl and with no SWC measures.Therefore Zn was spotted as the best micronutrient differentiating SWC measures.These differences can be explained by the influences of the tested SWC measures.Kyaruzi [7] in Usambara Mountains, also reported bench terraces and grass strips to have an influence on soil chemical properties such as pH, total N, OC, CEC, Ca 2+ and Mg 2+ when compared to control.Similar observations were reported by Tenge [9] and Wickama et al. [25] in Usambara Mountains, where soil conservations measures suchas bench terraces, Fanya Juu terraces and grass strips were found to have a big influence on soil chemical and physical properties as compared with cropland with no SWC measures.
On the other hand soil physical properties were significantly (P = .05)different between SWC measures except for soil texture which did not differ (Table 3).The available moisture contents (AMC) were higher under miraba with mulching than under miraba sole and cropland with no SWC measures.Bulk density (BD) values were significantly lower under miraba with mulching than under miraba sole and cropland with no SWC measures.Thus AMC and BD were powerful soil physical properties that discriminated SWC measures.The higher AMC and lower bulk density under miraba with mulching can be explained by the increased organic carbon contents due to the application of mulches (Tables 2 & 3).It has been established that the higher the organic carbon contents in the soil the lower the bulk density while also the higher the capacity of the soil to retain moisture available to plants [26].

The Influence of Selected SWC Practices on Crop Yields in Majulai and Migambo Villages
Maize and bean yields under the studied SWC practices in the Majulai and Migambo villages are presented in Table 4. Significant (P = .05)differences in crop yields between SWC practices were observed.Maize and bean grain yields followed the trend: miraba with Tughutu > miraba with Tithonia > miraba sole > control in both villages (Table 4).Maize grain yields were significantly (P = .05)higher in 2013 than in 2012, but there were no significant (P = .05)differences in bean grain yields between the two years of study.It was clearly observed that crop yield differences between treatments were highly influenced by the SWC practices (Table 4), while the higher crop yields under miraba with Tithonia and miraba with Tughutu mulches could be explained by the improved soil properties especially of AMC, OC, N, P, K, Ca 2+ , Mg 2+ , pH and BD (Tables 2 & 3).Similar observations were reported by Tenge [9] where Fanya Juu terraces had significantly higher maize and bean yields than under bench terraces and grass strips while control was the least; likewise the study by Msita [10] found miraba with farmyard manure and mulching to have higher maize and bean yields than miraba sole and control had the least.The higher yields were associated with improved soil fertility.The observed crop yields under the studied SWC practices (Table 4) were higher than the average yields according to FAO [29] of 1.5 Mg ha -1 for maize and of 0.7 Mg ha -1 for beans in Tanzania.
When considering variability of crop yields within the studied SWC practices, it can be seen from Table 4 that, crop gra in yields did not significantly (P = .05)varied within SWC measures except under cropland with no SWC measures where lower segments had higher maize grain yields than the upper segments.It can easily be noted that maize crop is more sensitive to the effect of gradients than bean crop; this is probably due to the ability of bean to fix nitrogen for its consumption as opposed to maize crop.Tenge [9] reported similar observations where bean crop performance was found not sensitive to slope gradients as opposed to maize.The evenly distributed crop yields within the studied SWC practices can partly be explained by the effect of reducing spacing of grass barriers that form miraba from the traditionally very wide to 5 m apart.This spacing was close enough to limit runoff velocity and thus reduced soil nutrients that could move with it down the slope to the lower segments.Besides, with this spacing, miraba were progressively forming bench terraces which cut down the slope and thus reduce translocation of soil nutrients by runoff.On the other hand mulching was also contributing to the reduced soil nutrient movement from the upper to the lower segments, allowing crops to respond evenly within the studied SWC practices.

Relation between Soil Properties and Crop Yields under the Different SWC Measures
Correlation between soil properties (that discriminated SWC measures) and crop yields are presented in Table 5.It can be seen that all the discriminator soil properties were positively correlated with crop yields except bulk density which was negatively correlated.The negative correlation of bulk density with crop yields can be explained by the fact that, bulk density is greatly influenced by soil organic carbon contents such that low the OC contents high the bulk density of the soils and vice versa (Table 2 & 3).Similar relationship was also reported by Aticho [26].Soil OC has been acknowledged to be an important cushion for many soil nutrients, thus the higher the OC content the higher the soil nutrients in the soil [23,24].A multiple linear regression model was fitted through the discriminate or soil properties that were correlated with crop yields under SWC measures (Table 6).It was found that maize grain yields were significantly (P = .05)a function of Ca   and K + (R 2 = 0.90).These observations imply that AMC and pH had greater potential of influencing maize and bean grain yields under miraba, while AMC, available P and K + had greater potential of influencing maize and bean grain yields under miraba with Tithonia and miraba with Tughutu mulching.The enhanced ability of miraba to avail soil water to plants and increase soil pH can be explained by the improved soil OC and exchangeable bases under miraba (Tables 2 &  3).Similar positive correlations of exchangeable bases with pH and AMC with OC were also reported by Mwango [23], Msanya et al. [24] and Shelukindo et al. [30].The improved P and K + were greatly due to the influences of mulching materials applied which have high contents of available P and K + (Table 1).This is strongly supported by the findings that applications of organic materials in soils reduce P sorption capacities and increase P availability [31], while also application of high quality organic materials with P content equal to or greater than 3.0 g P kg -1 in the soil decreases P adsorption [32], a tendency that improves P availability in the soil.

CONCLUSION AND RECOMMENDATIONS
Most of the studied chemical and physical soil properties were significantly (P = .05)influenced by the studied SWC measures.The trend for total N, OC, available P, Ca 2+ , Mg 2+ , K + and pH was: miraba with Tughutu > miraba with Tithonia > miraba sole > cropl and with no SWC measures (Control), while Na + did not differ.Micro nutrients Fe and Cu did not differ between SWC measures except for Zn and Mn which were significantly (P = .05)low in cropl and with no SWC measures.Likewise, miraba with Tughutu mulching had the highest AMC and lowest BD, whereas cropland with no SWC measures had the lowest AMC and highest BD.Maize and bean grain yields differed significantly (P = .05)inthe following trend: miraba with Tughutu > miraba with Tithonia > miraba sole > control in both villages.Crop grain yields did not significantly (P = .05)varied within SWC measures except for control which had higher crop grain yields in the lower segments than the upper segments.AMC and pH had the greatest potential in influencing maize and bean grain yields under miraba, while AMC, available P and K + had the greatest potential in influencing maize and bean grain yields under miraba with Tithonia or miraba with Tughutu mulching.Further researches are recommended to investigate the potentials of these mulching materials and their influences for the production of vegetables such as cabbage, tomatoes, onions and carrots which are widely cultivated in the Usambara Mountains.

Table 2 . The influence of the studied SWC practices on soil chemical properties Village SWC N pH OC % N % P K + Ca 2+ Mg + Na + Fe Mn Zn Cu Mg kg -1 cmol (+) kg -1 Mg kg -1
LSD: least significant different; SE: standard error of means

Table 3 . The influence of the studied SWC practices on soil physical properties
LSD: least significant different; SE: standard error of means