Effect of tillage system and nitrogen fertilization on the pH, extractable phosphorus and exchangeable potassium of Nitisols in Western Ethiopia

Maize production in Western Ethiopia is constrained by non-sustainable cropping practices, particularly repeated conventional tillage and insufficient N fertilization. A change in these cropping practices have usually a profound influence on soil chemical properties of which some may affect crop performance. Field trials were therefore conducted to determine the integrated effects of tillage system and nitrogen fertilization on the pH, extractable phosphorus and exchangeable K of Nitisols at five sites using maize as test crop from 2000 to 2004 in Western Ethiopia. Three tillage systems (MTRR = minimum tillage with residue retention, MTRV = minimum tillage with residue removal and CT = conventional tillage) and three N levels (the recommended rate and 25% less and 25% more than this rate) were combined in factorial arrangement with three replications. After five years the influence of the tillage systems on pH, extractable P and exchangeable K was confined to the upper 0-7.5 cm. The soil was acidified much more with MTRR than with either MTRV or CT. However, MTRR resulted in higher contents of extractable P and exchangeable K than MTRV and CT. Application of N fertilization for five consecutive years significantly decreased pH irrespective of tillage system. Neither extractable P nor exchangeable K were affected by N fertilization. The enhanced soil acidification that coincides with MTRR may impact negatively on maize production in the long run without a proper liming program. Liming is currently not a common practice in Western Ethiopia because of resource poor farmers. This aspect should be taken into account when the replacement of CT with MTRR is recommended. The findings of this study could be useful to other highland regions in Africa where cropping on Nitisols is common.


INTRODUCTION
The western part of Ethiopia has a high maize production potential because of favourable environmental conditions.However, maize production is constrained by nonsustainable cropping practices, particularly plough-or hoe-based cultivation (Bezuayehu et al., 2002), soil and water loss due to erosion, as well as N deficiency (Tolessa et al., 2002).This necessitated a proper investigation into the integrated effects of tillage system and N fertilization on the performance of the maize crop and the change in soil properties.The ultimate aim with this investigation was to obtain substantiated information whether the sustainability of the current cropping systems can be improved with minimum tillage.Results of the investigation concerning effects on yield and yield components (Tolessa et al., 2007) and efficacy of applied nitrogen (Tolessa et al., 2009) have already been reported but not that of changes in soil fertility indicators.
Tillage systems have a profound effect on the chemical properties of soils that ultimately may influence crop growth and development.It is especially pH and content of immobile nutrients such as P and K that are substantially affected by different tillage systems (Juo and Lal, 1979;Dick, 1983;Ismail et al., 1994;Kotzé and Du Preez, 2008;Lopex -Fando and Pardo, 2009).
The pH in the upper few centimeters of a soil usually decreases rapidly under minimum tillage compared to conventional tillage, especially when high rates of N fertilizer are used (Moschler et al., 1973;Blevins et al., 1983;White, 1990;Yadvinder-Singh et al., 2005).This drop in pH is attributed mainly to the H + released through the nitrification of NH 4 + .The NH 4 + originates from the surface-applied nitrogenous fertilizers and the N mineralized from the crop residues (Ismail et al., 1994).Organic acids released during the breakdown of crop residues could also contribute to the pH reduction (Stevenson and Cole, 1999).Weil and Magdoff (2004), however, pointed out that the organic matter which usually accumulates near the surface of minimum tilled soils tend to ameliorate the acidity.Thus, changes in pH resulting from tillage systems may have a bearing on fertilizer application strategies.
Accordingly, it has been found that minimum tillage in comparison with conventional tillage increased the concentration of plant nutrients like P and K in the surface layers of soil but decreased concentrations in deeper soil layers (Shear and Moschler, 1969;Lal, 1976;Juo and Lal, 1979;Ismail et al., 1994;Du Preez et al., 2001;Wright et al., 2007).Robbins and Voss (1991) for example reported that extractable P and exchangeable K in the 0-5 cm layer of minimum tilled soil were on average 3.5 times higher than in the 5-15 cm layer.Thus it is not surprising that these two plant nutrients are usually higher in the surface layers of minimum tilled soil than in conventionally tilled soil.For example, in a 20-year old trial it was found that extractable P and exchangeable K in the 0-5 cm layer of minimum tilled soil exceeded those in conventionally tilled soil by 1.5 and 1.8 times, respectively (Kotzé and Du Preez, 2008).This stratification of the two nutrients on account of minimum tillage can be attributed inter alia to: (1) no incorporation of surface-applied fertilizer P and K into soil; (2) uneven extraction of soil P and K by crop roots; (3) release of plant P and K from crop residues that decompose on the soil surface; and (4) little movement of P and K in soil due to their immobility (Dick, 1983;Unger, 1991;Lupwayi et al., 2006).
Stratification of other nutrients such as Ca, Mg, Cu, Fe Mn and Zn has also been associated with the adoption of minimum tillage (Lal, 1976;Blevins et al., 1983;Shuman and Hargrove, 1985).However, in a few instances no stratification of P (Juo and Lal, 1979;Lal, 1997), K (Hargrove et al., 1982) and Ca (Blevins et al., 1977) were found with minimum tillage.Reasons for these exceptions to the rule are given by the aforementioned authors.
The growth and development of a crop can be affected if the introduction of a tillage system results in a severe decrease of pH and increase of either extractable P or exchangeable K in the surface layers of soil (Matawo et al., 1999).Quantification of these processes are therefore of great importance in the evaluation of a tillage system before being advocated.
We are not aware of any research in Western Ethiopia on the temporal and spatial dynamics of soil fertility indicators for maize production under different tillage systems.Due to a lack of inter alia such information neither extension officers recommend nor smallholder famers apply conservation agriculture practices like minimum tillage generally.Thus it is not surprising that the sustainability of maize production in Western Ethiopia is questioned.This scenario can be reversed when sound experimental knowledge is available concerning changes in soil fertility indicators that coincide with the introduction of conservation tillage systems.The temporal and spatial effects of tillage system and N fertilization on the pH, extractable P and exchangeable K of Nitisols in Western Ethiopia are therefore reported in this paper.

Experimental sites
The field trials for this study were conducted under rainfed conditions at Bako Agricultural Research Center, and on farmers' fields at Shoboka, Tibe, Ijaji and Gudar.These five sites were selected to be representative of the major maize producing areas of Western Ethiopia in terms of climate and soil.Bako is located at 09° 01'N and 37° 02'E, Shoboka at 09°06' N and 37°21'E, Tibe at 09°29'N and 37°32'E, Ijaji at 09°43'N and 37°47'E, and Gudar at 08°09'N and 38°08'E latitude and longitude, respectively.
The altitudes for Bako, Shoboka, Tibe, Ijaji and Gudar are 1650, 1695, 1730, 1820 and 2000 m above sea level, respectively.Only climatic data of the Bako site with the lowest altitude and the Gudar site with the highest altitude as obtained from nearby weather stations is given in Table 1 since there are no weather stations close to the other three sites.Based on these data the mean annual rainfall over a 15-year period (1990)(1991)(1992)(1993)(1994)(1995)(1996)(1997)(1998)(1999)(2000)(2001)(2002)(2003)(2004)) ranged from 1042.2 mm at the higher lying Gudar site to 1239.6 mm at the lower lying Bako site, viz. a difference of 197.4 mm.For the cropping season (May to October) the average minimum temperature was 3.5°C lower and maximum temperature 0.9°C higher at the Bako site compared to that of the Gudar site.At all five sites the soil was classified as a Nitisol (FAO, 1998).Some physical and chemical topsoil characteristics of these Nitisols before commencement of the trials are summarized in Table 2.
The textural class of the Nitisols differed from loam at the Ijaji site to clay at the Shoboka site.Similar differences of 0.61 units in pH, 1.08% in organic C, 0.04% in total N, 3.9 mg kg -1 in extractable P and 85 mg kg -1 in exchangeable K were recorded between the five sites.The aforementioned differences in climate and soil justified therefore the selection of the five sites for this investigation.-------%--------------%-----------mg

Field trial layout
At each of the five sites a field trial was laid out in a randomized complete block design.The layout consisted of two factors namely, three tillage systems (MTRR = minimum tillage with residue retention, MTRV = minimum tillage with residue removal, and CT = conventional tillage) and three N fertilization levels (69 kg ha -1 , 92 kg ha -1 , and 115 kg ha -1 ) replicated three times in a complete factorial combination.Every field trial had therefore 27 plots.An application of 92 kg N ha -1 is the recommended fertilization rate for conventional maize production at the study sites, implicating the two other rates are 25% less and 25% more than this recommended rate.These experiments were conducted from 2000 until 2004.The experimental plots were kept permanent to observe the carry-over effects of the treatments for the five cropping seasons.

Agronomic practices
Before initiation of the trials the fields at all sites were under conventional maize production for many years.During the entire trial period immediately after harvesting the plants were cut at ground level and uniformly spread on the CT and MTRR plots, and removed from the MTRV plots.For the MTRR and MTRV treatments soil disturbance was restricted to the absolute minimum, viz. the soil was disturbed only to place the seed in the soil at the time of sowing.In contrast, the soil was ploughed three times with the local oxen-plough 'maresha' prior to sowing to obtain a suitable seedbed for the CT treatments.Urea and triple super phosphate were used as the sources of N and P, respectively.The application of urea was split and therefore half of the urea and all of the triple super phosphate were band placed 5 cm below the seed at sowing.At 35 days after sowing when maize was at knee-height the other half of the urea was band placed next to the row at 5 cm depth.The fertilizer in the small furrows was covered with soil soon after application.All treatments received the recommended phosphorus rate of 20 kg ha -1 annually.
Weed control in the MTRR and MTRV treatments was done by applying Round-up (glyphosate-isopropylamine 360 g a.i.L -1 ) at a rate of 3 L ha -1 prior to planting and Lasso/Atrazine (alachlor/atrazine 336/144 g a.i.L -1 ) at a rate of 5 L ha -1 as a preemergence application.The recommended weed control practice for CT in Ethiopia is hand weeding at 30 and 55 days after sowing followed by slashing at milk stage.
The standard cultural practices as commonly recommended to the farmers were adopted for the study.Therefore, from 2000 to 2004 the planting dates varied from 5 May to 5 June at all the sites.A late maturing commercial maize hybrid, BH-660 was planted.The plant density aimed for was 50 000 plants ha -1 as the 5.0 x 4.8 m plots consisted each of six rows, 5.0 m in length and the inter-and intra-row spacing was 0.8 and 0.25 m, respectively.

Data collection
Soil samples were collected, just before the trials commenced, from the 0-30 cm layer of all five sites for their characterization.A 5 cm diameter auger was used to sample 20 randomly selected spots per site.These subsamples were thoroughly mixed, dried at room temperature, sieved through a 2 mm screen and stored until analysis.Since the trials started soil samples were collected annually after harvesting from the 0-30 cm layer of all plots at each site.At the end of the trial period the 0-7.5 cm, 7.5-15 cm, 15-22.5 cm and 22.5-30 cm layers were sampled additionally.In both instances an auger with a 2 cm diameter was used to sample five randomly selected spots per plot.
These subsamples were prepared for analysis as described earlier.Standard procedures (The Non-affiliated Soil Analysis Work Committee, 1990) were used to determine particle size distribution (Hydrometer), organic C (Walkley-Black), total N (Kjeldahl), pH (1:2.5 water), extractable P (Bray 2) and exchangeable K (NH 4 OAc) of the relevant composite soil samples.

Statistical analysis
Experimental data were analyzed through analyses of variance using the MSTATC statistical package (Michigan State University, 1989).Means for each parameter were compared by the least significant difference (LSD) test at P = 0.05.

RESULTS AND DISCUSSION pH
Both tillage system and N fertilization had a significant effect on the pH of the 0-30 cm soil layer for the year 2000 to 2004 but there was no significant interaction between the two main effects.The effect of tillage system on pH in this soil layer is illustrated in Figure 1.At all five sites pH decreased over the experimental period regardless of the tillage system applied.This decrease in pH was least severe with CT, followed by MTRV and then MTRR.Differences in pH between tillage systems were initially small and inconsistent but as the experiments progressed it become more apparent.The magnitude of pH differences between CT and MTRR was surprisingly similar for the five sites, namely in the order of 0.07 units.
The pH of the 0-30 cm soil layer at all five sites decreased from 2000 to 2004 irrespective of the N application level (Figure 2).This decrease was least severe with the 69 kg ha -1 N level, followed by the 92 kg ha -1 N level and then the 115 kg ha -1 N level.The differences in pH between the three N application rates were only significant from the year 2003.This is probably due to the accumulative effect of the H + released with the continued use of urea at the three fixed rates.In the year 2004 pH differed between the lowest and highest N application level from 0.08 units at the Tibe site to 0.14 units at the Ijaji site.This is an indication that the buffer capacity of the Nitisols in this investigation was not similar.
As shown in Figure 3, the pH differences that evolved in the upper 30 cm of the Nitisols from either tillage system or N fertilization are attributable to their effects in the 0-7.5 cm layer.In this layer the highest pH was recorded in the CT soil, followed by the MTRV soil and then the MTRR soil.At the five sites the difference in pH between the CT and MTRR soils ranged from 0.15 to 0.25 units.The pH of the next three soil layers was not affected significantly by the three tillage systems at any of the sites.
The increase of pH with depth is common in the Nitisols of the study area.However, acidification of the upper 7.5 cm of these soils at all the sites appeared to be occurring faster with MTRR than with MTRV or CT.This phenomenon could be attributed to the nitrification of NH 4 + released from either the fertilizer or residues at or near the soil surface (Blevins et al., 1983;Ismail et al., 1994) since the process produces acidifying hydrogen ions (Fox and Bandel, 1986).Similar changes in pH on account of tillage systems were reported by other researchers (Shear and Moschler, 1969;Blevins et al., 1983;White, 1990;Du Preez et al., 2001;Kotzé and Du Preez, 2008).

Extractable P
Extractable P of the 0-30 cm soil layer was significantly affected by tillage system and not by N fertilization.The effect of tillage system on extractable P in this soil layer is displayed in Figure 4 for the year 2000 to 2004.Regardless of year MTRR exhibited a higher extractable P level than MTRV while CT was intermediate.These differences in extractable P levels between tillage systems were only significant from 2002 at Gudar, 2003 at Baka and Ijaji, and 2004 at Shoboka.However, the differences of extractable P between MTRR and MTRV soils at these four sites were relatively small in the year 2004, ranging from 0.9 mg kg -1 at Gudar to 1.3 mg kg -1 at Ijaji.Differences of this magnitude should not have any influence on the growth and development of maize.
The above mentioned differences in extractable P originated in the upper 15 cm soil layer as no significant differences were found below this depth (Figure 5).In 2004 irrespective of site, the extractable P level of the 0-7.5 cm layer was significantly higher in the MTRR soil than in either the MTRV or CT soils.The differences ranged from 2.5 mg kg -1 at Tibe to 5.1 mg kg -1 at Gudar which are probably large enough to affect maize growth and development.However, the extractable P level of the 7.5-15 cm layer was higher in the CT soil than in either the MTRV or MTRR soils although not significant at all sites.
The higher extractable P levels recorded especially in the 0-7.5 cm soil layer and to a lesser extent in the 7.5-15 cm layer of the Nitisols regardless of the tillage system can be attributed to the immobility of this nutrient.However, as indicated the tillage systems caused after five consecutive years of practice different extractable P levels in the upper 7.5 cm layer.The higher extractable P levels in this layer of the MTRR than the CT soils can be attributed to the applied P fertilizer and the retained maize residues which were not mixed with the soil to the same degree due to the nature of the two tillage systems.It seems however that the retention of maize residues contributed largely to this phenomenon as the extractable P levels in the upper 7.5 cm layer of the MTRV and CT soils were almost similar.The retained maize residues on the soil surface enhanced organic matter formation and in this process some of the P taken up by the crop from deeper layers is released in an inorganic form (Ismail et al., 1994;Thomas et al., 2007).This released inorganic P is probably less subject to fixation as organic matter can protected it to some degree (El-Baruni and Olsen, 1979).

Exchangeable K
Exchangeable K of the 0-30 cm layer was significantly affected by tillage system and not by N fertilization.The effect of tillage system on exchangeable K in this soil layer is shown in Figure 6 for the year 2000 to 2004.
During the first three years exchangeable K differed not significantly among the three tillage systems which was not the case in the last two years at Bako, Shoboka, and Tibe when significantly higher levels of exchangeable K were recorded in the MTRR than MTRV and CT soils.However, throughout the experimental period the MTRR soils and to a lesser extent also the CT soils exhibited higher levels of exchangeable K than the MTRV soils.Thus at termination of the trials at the five sites in the year 2004 exchangeable K averaged 176 mg kg -1 for MTRR soils, 169 mg kg -1 for the CT soils, and 139 mg kg - 1 for the MTRV soils.
The above mentioned differences in the exchangeable K originated in the upper 15 cm soil as no significant differences were recorded below this depth (Figure 7).In 2004 after five consecutive years of practice, MTRR resulted in the highest exchangeable K level in the 0-7.5 cm soil layer at all sites, followed by MTRV and CT.The exchangeable K levels for this layer averaged therefore over the five sites at 233 mg kg -1 in the MTRR soils, 202 mg kg -1 in the CT soils, and 187 mg kg -1 in the MTRV soils.However, in the 7.5-15 cm soil layer at all sites, the exchangeable K levels MTRR and CT were almost similar but higher than that of MTRV.
A decline of exchangeable K with depth in the Nitisols is common to the study area.However, the differences in exchangeable K that evolved in the upper 15 cm of the Nitisols on account of tillage systems are a consequence of the concomitant residue management since no K fertilizer was applied.Several researchers (Triplett and Van Doren, 1969;Fink and Wesley, 1974;Addiscott and Dexter, 1994;Du Preez et al., 2001;Ben Moussa-Machraoui et al., 2010) showed that the fate of maize residues had a large influence on exchangeable K in soils as the residues contain a large amount of K.The decrease of pH and increase of extractable P and exchangeable K in the 0-7.5 cm layer of MTRR soils with respect to MTRV and CT soils had probably no negative effect on the grain yield of maize during the final two years of the trial period.In this period there was no significant difference in grain yield between MTRV and CT and both were significantly inferior to MTRR (Tolessa et al., 2007).However, it can be expected that the changes of pH, extractable P and exchangeable K resulting in the Nitisols within five years of MTRR will proceed with continuous application of this tillage system.Prolonged application of MTRR may have eventually a negative impact on the performance of the maize crop, especially through enhanced soil acidification.The rectifying of soil acidity through liming is currently not a common practice in Western Ethiopia because most of the resource poor farmers cannot afford it.This aspect should be taken into account when the replacement of CT with MTRR is advocated.Compared to CT, MTRR resulted in more efficient use of applied N by maize (Tolessa et al., 2009).
We are of opinion that the findings of this study will be applicable also to the remaining highland regions of Ethiopia as well as those of other African countries like Cameroon, Congo and Kenia.In these countries are cropping on Nitisols common, notably at altitudes of more than 1200 m above sea level.The estimated area of Nitisols in the higlands of Africa is 100 million ha.

Conclusions
Tillage systems and concomitant residue management showed profound effects on the pH, extractable P and exchangeable K in especially the 0-7.5 cm layer of the studied soils.Acidification was more pronounced with MTRR than with MTRV and CT.This was also found with an increase of N fertilization from 69 to 115 kg ha -1 .The extractable P and exchangeable K contents of the MTRR soils were higher than that of the CT and MTRV soils.Throughout the five year trial period MTRR at least maintained the extractable P and exchangeable K contents, while with MTRV the contents of both these nutrients decreased steadily.The decline in extractable P and exchangeable K with prolonged MTRV are indicative of soil fertility degradation.Hence, soil fertility can be maintained or restored for sustainable crop production in Western Ethiopia with MTRR if care is taken of the acidification that coincide with this tillage system.This study's findings could be extrapolated to the remaining highland regions of Ethiopia as well as those of other African countries where cropping is practiced on Nitisols.

Figure 1 .
Figure 1.Effect of tillage system on pH measured after harvesting in the 0-30 cm layer of Nitisols at the five experimental sites in 2000 to 2004.

Figure 2 .
Figure 2. Effect of nitrogen fertilization on pH measured after harvesting in the 0-30 cm layer of Nitisols at the five experimental sites in 2000 to 2004.

Figure 3 .
Figure3.Effect of tillage system on pH measured after harvesting in the 0-7.5 cm, 7.5-15 cm, 15-22.5 cm and 22.5-30 cm layers of Nitisols at the five experimental sites in 2004.

Figure 4 .
Figure 4. Effect of tillage system on extractable P measured after harvesting in the 0-30 cm layer of Nitisols at the five experimental sites in 2004.

Figure 5 .
Figure5.Effect of tillage system on extractable P measured after harvesting in the 0-7.5 cm, 7.5-15 cm, 15-22.5 cm and 22.5-30 cm layers of Nitisols at the five experimental sites in 2004.

Figure 6 .
Figure 6.Effect of tillage system on exchangeable K measured after harvesting in the 0-30 cm layer of Nitisols at the five experimental sites in 2000 to 2004.

Figure 7 .
Figure7.Effect of tillage system on exchangeable K measured after harvesting in the 0-7.5 cm, 7.5-15 cm, 15-22.5 cm and 22.5-30 cm layers of Nitisols at the five experimental sites in 2004.

Table 1 .
Climatic data for the Bako and Gudar sites as obtained from nearby weather stations.

Table 2 .
Some physical and chemical topsoil characteristics of the Nitisols at the study sites before commencement of the trials.