Variation in Soil Properties under Long-Term Irrigated and Non-Irrigated Cropping and Other Land-Use Systems in Dura Catchment, Northern Ethiopia

There are limited reports about the impacts of long-term irrigated and non-irrigated cropping and land-use systems (CLUS) on soil properties and nutrient stocks under smallholder farmers’ conditions in developing countries. The objective of this research was to examine variation in soil properties and OC and TN stocks across the different CLUS in Dura sub-catchment, northern Ethiopia. Surveys and discussions on field history were used to identify nine CLUS, namely, tef (Eragrostis tef (Zucc) Trot)) mono-cropping (TM), maize (Zea mays L.) mono-cropping (MM), cauliflower (Brassica oleracea var. botrytis)-maize intercropping (IC1), red beet (Beta Vulgaris)-maize intercropping (IC2), cauliflower-tef-maize rotation (R1), onion (Allium cepa L.)-maize-onion rotation (R2), treated gully (TG), untreated gully (UTG), and natural forest system (NF). A total of 27 composite soil samples were collected randomly from the CLUS for laboratory analysis. Data were subjected to one-way analysis of variance and PCA. The lowest and highest bulk density was determined from NF (1.19 Mg m-3) and UTG (1.77 Mg m-3), respectively. Soil pH, EC and CEC varied significantly among the CLUS. The highest CEC (50.3 cmolc kg-1) was under TG followed by NF. The highest soil OC stock (113.6 Mg C ha-1) and TN stock (12.2 Mg C ha-1) were found from NF. The PCA chosen soil properties explained 87% of the soil quality variability among the CLUS. Such soil properties and nutrient stocks variability among the CLUS suggested that introduction of suitable management practices are crucial for sustaining the soil system of the other CLUS.


51
Soil quality is becoming an important resource to raise crop productivity so that to meet the food 52 required for the current and future population in developing countries as their economy mainly 53 depends on agriculture [1][2][3][4][5][6]. Soil quality is defined as the capacity of the soil to give the 54 intended functions for biomass and yield production [7][8][9]. In this study, the term soil quality is 55 used synonymously with soil health. Recently, however, soil quality degradation caused by 56 inappropriate cropping system, and land-use and soil management practices, has been reported information on site-specific soil properties is a basic tool for proper soil management in order to 115 provide sustainable soil functions at present and in the future [2,4,14,25]. Site-specific data on 116 soil properties could also support to deal with spatial variability of soil nutrients and physical 117 indicators and their influencing factors. Such information is important to formulate appropriate 118 sustainable cropping systems and land-use type strategies [6,12,14,19]. 119 The sustainability of crop and soil management practices to improve or maintain soil quality 120 depends on understanding how soils respond to different site-specific cropping and land-use   In Dura sub-catchment, both rain-fed and irrigation agriculture have been practiced for more 156 than 2 decades. But rain-fed agriculture which is the oldest practice dominated in area coverage. 157 About 100 ha farmland has been irrigated since 1996 in Dura sub-catchment. Afforested area, 158 pasture, scattered woody trees, bushes and shrub lands were also found in the study sub-  cropping and land-use systems. The purpose of the reconnaissance survey was to characterize 172 the fields' historical cropping system, soil management, agronomic practices and field features.

173
During the survey, participatory tools such as field observation, transect-walks and group 174 discussion were employed. The transect-walks were done twice, that is, from the east to the 175 west and also from the north to south direction of the study sub-catchment in order to observe 7 176 different cropping and land-use systems. This was done by the team composed of the 177 researcher, three (3) DAs and randomly selected 10 farmers from the study sub-catchment. 178 Three group discussions sessions were held in order to reach consensus among the 179 participants about the descriptions of the irrigated and non-irrigated fields that were selected. 180 On the basis of the farmers' final consensus nine (9) dominant cropping and land-use systems 181 (fields) were identified (Table 1), and geo-referenced and described their topographic features 182 (Table 2). Such fields were selected because the site and crop specific management practices 183 perhaps affect the sustainability of natural resources, crop productivity and soil fertility 184 utilization in the catchment. The selected fields were located on Chromic Vertisols adjacent to 185 each other at a distance that ranges between 50 and 150 m.

186
Insert Table 1 here 187 From the total nine (9) fields identified, four (4) were from irrigated fields, two (2) from rain-  Table 1). The first two (TM and MM) were selected from the rain-fed fields 195 adjacent to the irrigation command area, whereas IC1, IC2, R1 and R2 were selected from the 196 irrigated crop fields. TG and UTG were also found within the irrigation command area. An 197 adjacent natural forest land-use system (NF) was used as a reference while compared with the 198 impact of irrigation and non-irrigation cropping and land-use systems on soil properties, and 199 carbon and nitrogen stocks.  expert knowledge is very efficient as it is quick and easy to select the sampling units.

212
Considering the costs of soil analysis and its statistical representativeness a total of 27 213 composite soil samples from three sampling units (9 CLUS x 3 sampling units) were collected.

214
The soil sampling unit plot area was 48 m 2 (6 m x 8 m). The soil sampling plots land features 215 are described in Table 2. In each sampling unit plot 10 pairs of randomly selected coordinate 216 points were identified. From the 10 geo-referenced points in each sampling plot, three sampling 217 points were selected using simple random sampling technique whereby the composite soil 218 sample from each plot was collected.

219
The soil samples were taken from each sampling point at 0-20 cm soil depth. This sampling 220 depth is where most soil changes are occurred due to long-term cropping systems, land-use 221 types, and soil and water management practices including irrigation agriculture. Three soil 222 samples were collected from each sampling unit in a plot and pooled (composited) into a bucket 223 and mixed thoroughly to form a single homogenized sample. A sub-sample of 500 g soil that 224 represented the pooled sample in the bucket was taken from each sampling unit plot, and air 225 dried and sieved through 2 mm mesh sieves. In addition, three undisturbed soil samples were 226 collected from each irrigated and non-irrigated field sampling unit plots at 0-20 cm soil depth 227 using 5·0 cm long by 5·0 cm diameter cylindrical metal core sampler to determine soil dry bulk 228 density.

229
The analysis of the soil samples was carried-out following the standard laboratory procedures  Soil pH was determined in 1:2·5 soil to water ratio using pH meter combined glass electrode 236 [49], electrical conductivity (EC) in 1:2·5 soil to water ratio using conductivity meter [ Where, SSSI (%) is soil structural stability index, SOC is soil organic carbon, and clay + silt 251 is combined clay and silt content. SSSI < 5% indicates structurally degraded soil; 5% < SSSI < 252 7% indicates high risk of soil structural degradation; 7% < SSSI < 9 % indicates low risk of soil 253 structural degradation; and SSSI > 9% indicates sufficient SOC to maintain the structural 254 stability. A higher the SSI value, a better would be in maintaining soil structural degradation.

255
Base saturation percentage was calculated by divided the sum of base forming cations (Ca 2+ , 256 Mg 2+ , K + and Na + ) by CEC and then multiplied by 100%. Exchangeable sodium percentage 257 (ESP) was calculated by divided exchangeable Na + by CEC. The ESP threshold of 15% was used 258 to classify sodium hazard, that is, sodic soils are those with ESP of more than 15%. Sodium 259 adsorption ratio (SAR) was calculated [58-60] as: Where, SAR is sodium adsorption ratio in (cmol kg -1 ) 0.5 ; and Na + , Mg 2+     statistically grouped into five principal components (PCs) using the Varimax rotation procedure.

296
The five PCs with eigenvalues > 1 that explained at least 5% of the variation of the soil 297 properties response to the cropping and land-use systems were considered. Varimax rotation with 298 Kaiser Normalization resulted in a factor pattern that highly loads into one factor [65]. If the 299 highly weighted variables within PC correlated at the correlation (r) value < 0.60, all variables 300 were retained in the PC. Among the well-correlated variables (r  0.60) within PC, a variable 301 with the highest partial correlation coefficient and factor loading was retained in the component Research Ethics Review Committee of Aksum University, Ethiopia to conduct this study. Full 314 right was given to the study participants to refuse and withdraw from participation at any time.

315
Confidentiality of respondents was preserved by the researchers during data collection of soil 316 samples and soil and crop management history. It was also noted that the research has no any 317 activities that directly related to human being as it is directly related to the physical environment.  (Table 3). The soil clay contents of the CLUS varied significantly between 26 to 74%, 326 with the lowest and the highest values observed from TM and R2, respectively. This could 327 influence the other textural classes and physical and chemical soil properties.

328
The lowest silt (22.7%) and sand (3.7%) contents were observed from R2 whereas the highest 329 silt (43%) from NF and sand (39.0%) from TM were observed. The highest sand content in the 330 TM may be associated with repeated cultivation using inorganic fertilizer for long-time in which 331 such practices aggravate erosion that erode fine soil particles and leaves coarser particles (Brady 332 and Weil, 2002). The mean clay (44.2%), silt (33.2%) and sand (22.6%) contents of all the 333 CLUS indicated that the sub-catchment soil has dominated by clay followed by silt texture soil.

334
Fields with higher clay content such as R2 are considered by local farmers as difficult for 335 workability and so susceptible to the problem of water logging in which this is in agreement with  On the other hand, there were non-significant differences in soil sand contents among some 338 of the CLUS, e.g., between TM and UTG, IC1 and IC2, R1 and TG (Table 3). This could be 339 attributed to the fact that sand texture is soil property that does influence little by some of the considered in this study showed DBD values higher than the critical level in which this implies 361 the need for adopting appropriate practices that improve soil bulk density.

362
Total porosity, SSSI and A-horizon values were significantly varied among the CLUS, with 363 the highest of these parameters recorded from NF and the lowest from UTG (Table 3). The trend 364 of these parameters is similar to that of DBD but in the opposite direction. The variation in total 365 porosity, SSSI and A-horizon among the different CLUS could be attributed to the differences in  There were statistically significant differences in soil chemical properties among most 389 cropping and land-use systems (CLUS) in the study sub-catchment (Table 4). The soil pH varied 390 significantly from 6.94 in TM to 8.50 in R1. The higher pH in R1 could be associated with the 391 effects of long-term irrigation and soil management practices. There was also non-significant 14 392 differences in soil pH among some of the CLUS (e.g., between MM and NF; and among IC1, 393 IC2, and TG). The mean soil pH (7.68) of all the CLUS indicates that the study catchment soil is 394 categorized as moderately alkaline in reference to the classification for African soils reported by 395 Landon [80]. Generally, the CLUS in the catchment showed soil pH values within the critical 396 levels (6.5-8.5) reported in literature (e.g., Sanchez  indicates that soil pH is not a key problem to monitor effects of the different cropping and land-398 use systems on soil quality indicators.

399
The highest soil EC was recorded from irrigated fields of IC2 (0.510 ds m -1 ) followed by R2 400 (0.390 ds m -1 ) whereas the lowest was found from rain-fed field of TM (0.057 ds m -1 ). However, 401 there were non-significant differences in EC among many of the CLUS (e.g., MM, IC1, R1, R2, 402 TG, UTG and NF) ( Table 4). According to Landon [80], soil EC determined from the different 403 CLUS is categorized as non-saline even though EC was found to be higher in the irrigation fields   The highest Ex Na was found from IC2 (0.682 cmol c kg -1 ) followed by IC2 (0.667 cmol c kg -430 1 ). The lowest Ex Na was recorded from NF (0.030 cmol c kg -1 ) and TM (0.040 cmol c kg -1 ). The 431 Ex Na recorded from TG was significantly higher than that of NF, TM, MM and UTG (Table 4), 432 in which this could be attributed to the effects of long-term irrigation water drained to TG as it is 433 located within the irrigation fields. According to Landon (1991), Ex Na observed from IC1, IC2, 434 R1, R2 and TG were rated as medium; MM and UTG as low; and NF and TM as very low.

435
However, the Ex Na from the irrigation fields was found to be near to the cut-off point for 436 medium rate (0.7 cmol c kg -1 ) which is regarded as potentially sodic, indicating that necessary soil 437 and crop management practices should be taken to reduce or maintain Ex Na of the soil. In 438 addition, the highest SBFC and BSP were found from NF and TG whereas the lowest was from 439 UTG followed by TM (Table 4). According to the report by Landon [80] the BSP from NF and 440 TG was rated as very high and that of UTG was rated as medium. According to the same author, 441 the BSP of the remaining CLUS were categorized in the high rate. The highest ESP and SAR 442 were recorded from IC1 followed by R2 and the lowest was from NF followed by TM (Table 4). 443 However, all the CLUS showed ESP < 2, which is classified as low or non-sodic soils as 444 reported in the rate for African soils by Landon (1991). Since the SAR is < 12 which is the cut-445 off point [58], the soil of the CLUS is categorized as non sodicity.  448 The highest and statistically significant Pav was recorded from NF (23.9 mg kg -1 ) followed by 449 TG (19.4 mg kg -1 ). However, the lowest Pav was found from UTG (1.4 mg kg -1 ) followed by 450 TM (2.0 mg kg -1 ). The Pav contents among the intercropping and crop rotation practices under 451 irrigation system were non-significantly differed. But Pav from MM was found to be 452 significantly higher than the other cropping systems (Table 4). Soil Pav of NF, TG, and MM 453 were rated as very high, high and medium, respectively, and the rest CLUS rated as very low in be given to CLUS with very low soil Pav so that to improve Pav using appropriate practices and 16 456 also maintain the Pav of potential land-use systems. The highest and statistically significant soil organic carbon (OC) was found from NF (4.98%) 463 followed by TG (3.120%) while the lowest OC was from UTG (0.413%) followed by TM 464 (0.643%). The optimal OC, i.e., between 3% < OC < 5 % as proposed by Craul [86], which  The highest soil total nitrogen (TN) was found due to NF (0.541%) followed by TG 487 (0.257%). The lowest TN was recorded from UTG (0.030%) followed by TM (0.067%). The soil 488 TN from MM was significantly higher than that of IC1, IC2, R1 and R2 (

501
The highest OC: TN was recorded from TG (13.7) followed by NF (12.9) and UTG (12.2) ( Table   502 4), which could be associated with low oxidation (decomposition) rate of organic sources as 503 compared to the inputs available in the study sub-catchment. Meaning, there were no soil and 504 agronomic practices that enhanced decomposition of organic sources in these selected land-use 505 systems. In addition, the soil in TG was water saturated almost for more than 8 months of the 506 year, in which this could be slow down the decomposition of organic matter by limiting soil 507 microbial activity [61,63]. Similarly, long-term effects of irrigation practices can reduce 508 microbial activity and thereby reduces organic matter decomposition which could be the reason 509 for OC: TN to be slightly higher than 10 in the irrigation fields such as IC1, IC2, R1 and R2. The 510 OC: TN of MM (11.4) was found to be higher than that of the fields under irrigation cropping     535 In the study sub-catchment, soil OC and TN stocks varied significantly among the majority of 536 the cropping and land-use systems at 0-20 cm depth (Table 5). The highest stock of OC (113.6 537 Mg C ha −1 ) and TN (12.2 Mg C ha −1 ) were reported from NF. The soil OC stock from NF was   PCs explained by each soil property ranges from 74-96% (Table 6). A high communality 594 variable shows that a high portion of variance was explained the variable and therefore, it gets a 595 higher preference to a low communality [100].

596
The highly loading variables in PC1 were CEC and clay content (Table 6). Since the 597 correlation coefficient between CEC and clay was 0.86 which is higher than the cut-off point 598 (0.60,), communality was considered to select the parameter to be retained in PC1. As a result, 599 CEC was retained in PC1 because the loading (0.87) and communality (0.95) of CEC were 600 higher than that of clay. The first PC is thus termed as 'cation exchange capacity, CEC factor'. that the contribution of TN and SSSI to PC2 is explained using SOC. As a result, PC2 is termed 607 as the 'organic matter factor'. The highly loaded variables in PC3 included dry soil bulk density 608 (DBD), porosity and A-horizon depth (A_h). The partial correlation analysis between DBD and 609 porosity showed at r = 0.85. Since the loading value and communality of DBD was slightly 610 higher than that of porosity, DBD was retained in PC3. A-horizon depth was also retained in PC3 611 as the correlation coefficient with the other high loading variables showed less than the cut-off 612 point (< 0.60). Thus, PC3 is termed as 'soil physical property factor'. The highly loaded variable 613 in PC4 included Ex Na, ESP and SAR. The partial correlation values among these variables 614 showed strong correlation (r > 0.88). Considering the higher correlation coefficient, loading 615 weight and communality values, Ex Na was retained in PC4 and the rest variables were excluded 21 616 from PC4. As a result, PC4 was termed as the 'sodicity factor'. Likewise, the highly loaded 617 variables in PC5 are TN and Pav (Table 6). Since the correlation between TN and Pav is below   Table 1. Cropping and land-use systems identified in the Dura sub-catchment, northern Ethiopia.    Continuous tef crop has been grown (mono-cropped) for more than 18-years at the same field. Inorganic fertilizer of 100 kg ha -1 DAP and 50 kg ha -1 urea applied in each crop seasons, but recently the application was 100 kg ha -1 of each these fertilizers. Tillage frequency to prepare seed bed ranges between three and six times, depending on field soil conditions and farmers resources availability. The tef fields are located near to the irrigated fields, but far from homesteads. Tef is rain-fed cropping system. Soil and water conservation (SWC) practices observed only at the boarder of the fields. No manure or compost fertilizer applied. Soil samples were collected just after harvested the tef crop.

935
2 Maize (Zea mays L.) mono-cropping (MM) Continuous maize crop has been grown for more than 30-years at the same field. Fields are located just at homesteads and received about 5 to 8 tones ha -1 of manure each year. Sometimes, 100 kg DAP ha -1 and 50 kg urea ha -1 were applied if organic sources are small or unavailable. Tillage frequency is at most three times. This is rain-fed cropping system, and relatively it has intensive SWC practices. Soil samples were collected just after harvested the maize crop. Intercropping of red beet with maize was practiced using furrow irrigation for a consecutive of five years. Maize was planted three weeks after red beet. Soil samples were collected at maturity stage of both crops. During the rain-fed season tef intercropped with "Nuhig" (Guizotia abyssinica L.) was used to rotate the system. Fertilizer rate of 100 kg DAP ha -1 and 50 kg urea ha -1 were applied during all cropping seasons. The land was ploughed three times. A small rate of manure/compost (1 ton ha -1 ) was applied occasionally just at planting of red beet.

5
Cauliflower -tefmaize rotation (R1) Irrigated (furrow irrigation) sole cauliflower was first planted. After this crop harvested, rain-fed tef was planted. After tef, irrigated maize was planted and then the rotation was continued to cauliflower followed by maize again for two terms (6 years). Soil samples were taken during the maturity stage of irrigated maize crop at the end of term two. The fertilizer rate applied for all of the crops included 100 kg DAP and 50 kg urea ha -1 . A small rate of manure/compost (1 ton ha -1 ) was applied occasionally just at planting time of cauliflower and maize.
6 Onion (Allium cepa L.) -maize -onion rotation (R2) Irrigated (furrow irrigation) onion was first planted. After this crop, rain-fed maize was planted as a rotational crop. After maize, irrigated onion was planted again and then continue the rotation to rain-fed maize and back to onion irrigation for two terms (6 years). Soil samples were taken during maturity stage of the irrigated onion at the end of the term two. The fertilizer rate applied to both crops included 100 kg DAP ha -1 and 50 kg urea ha -1 . A small rate of manure/compost (1 ton ha -1 ) was applied occasionally just at planting of irrigated onion.

7
Treated gully (TG) The gully in the irrigation command area was treated using Sesbania (Sesbania sesban) and Leucena (Leuceana leucacephala) legume trees which established 20-years ago. Naturally regenerated grasses have also grown well on the bed and sides of the gully and have used only by cut and carrying system. The gully treatment has entirely dependent on biological SWC. Excess irrigation water from the fields was drained to the treated gully.

Untreated gully (UTG)
The untreated gully had no improved management practices, e.g., no SWC, no enrichment of tree, shrub and grass species. This land has not been contributed to the local community livelihood for many years. According to local farmers, the estimated age of gully is more than 100 years.