Exploring micro-field water-harvesting farming system in dryland wheat (Triticum aestivum L.): An innovative management for semiarid Kenya

G Model

ARTICLE IN PRESS

FIELD-6745; No. of Pages 12

Field Crops Research xxx (2016) xxx–xxx

Contents lists available at ScienceDirect

Field Crops Research journal homepage: www.elsevier.com/locate/fcr

Exploring micro-field water-harvesting farming system in dryland wheat (Triticum aestivum L.): An innovative management for semiarid Kenya Jian-Yong Wang a,1 , Fei Mo a,1 , Simon N. Nguluu b , Hong Zhou a , Hong-Xu Ren c , Jian Zhang a , Charles W. Kariuki b , Patric Gicheru b , Levis Kavaji d , You-Cai Xiong a,∗ , Feng-Min Li a a

State Key Laboratory of Grassland Agro-ecosystems, Institute of Arid Agroecology, School of Life Sciences, Lanzhou University, Lanzhou 730000, China Kenya Agricultural and Livestock Research Organization, Kabete, 14733-00800, Nairobi, Kenya c The Institute of Botany, Chinese Academy of Sciences, Xiangshan, Beijing 100093, China d United Nations Environment Programme, P.O. Box 47074-00100, Nairobi, Kenya b

a r t i c l e

i n f o

Article history: Received 17 March 2016 Received in revised form 27 June 2016 Accepted 2 July 2016 Available online xxx Keywords: Micro-field rain-harvesting farming system Productivity Profitability Dryland wheat Semiarid Kenya

a b s t r a c t Micro-field rain-harvesting farming system (MRFS) has demonstrated great potentials to enhance field productivity and profitability of dryland wheat (Triticum aestivum L.) in semiarid eastern Asia, yet little is known whether this system results in desired effects in semiarid Africa such as Kenya. A two-year field experiment was conducted during 2012 and 2013 growing seasons to evaluate the effects of introduced MRFS on water availability, field productivity and economic benefits using a local wheat (T. aestivum L.) cultivar DUMA in a semiarid site of Kenya. Five treatments were designed as: 1) ridge and furrow with transparent plastic mulching (RFT); 2) ridge and furrow with black plastic mulching (RFB); 3) ridge and furrow with grass straw mulching (RFS); 4) ridge and furrow without mulching (RF); and 5) traditional flat planting (CK). The results showed that mulching treatments (RFT, RFB and RFS) significantly decreased the evapotranspiration (ET) by 11.4–88.5 mm, increased wheat grain yield by 60%–163%, above-ground biomass by 58%–104% and water use efficiency for grain by 68%–271%, compared with CK over two growing seasons. RFT and RFB treatments resulted in maximal soil water storage at 1-m depth and the greatest harvest index among all treatments. Linkage analyses indicated that grain yield showed significantly positive correlation with plant height, leaf area and major spike components (P ≤ 0.05), suggesting that plant type of wheat was altered for better yield production as a result of MRFS operation. More importantly, economic ratios of output to input were also calculated and compared. The average ratio of output to input for CK was 3.86, slightly lower than 4.28, 4.06, 5.86 and 5.34 for RFT, RFB, RFS and RF, respectively across two growing seasons. In particular, net incomes in MRFS (RFT, RFB, RFS and RF) were increased by 145%, 128%, 117% and 82% respectively, compared with that of CK. In conclusion, on-field rain-harvesting farming system provides an innovative management to boost the productivity and profitability of dryland wheat, and a potential solution to cope with food security in semiarid Kenya. © 2016 Elsevier B.V. All rights reserved.

1. Introduction Adaptive management of agricultural system has become a hot issue under climate change and food security (Klerkx et al., 2010; Lehmann et al., 2013; Rochecouste et al., 2015). Climate change brings about enormous challenges to agricultural production and food security, particularly in developing countries and

∗ Corresponding author. E-mail addresses: [email protected], [email protected] (Y.-C. Xiong). 1 Jian-Yong Wang and Fei Mo contribute equally to this work.

rain-fed areas (Beddington et al., 2012; Garnett et al., 2013). It will also continue to affect global precipitation and evaporation. In rain-fed agricultural areas where the irrigation and groundwater resources are not available, low-moisture and variable soil environment exerts increasing negative impacts on crop production (Morton, 2007; Townes, 2015). Low productivity of dryland agriculture leads to serious food shortage and affects livelihood development. To date, more than 2 billion people in developing countries survive on less than 2 US dollars per day, and spend most of it on food. Most of these people live in drought-prone areas, including Kenya and other areas of Africa. Kenya belongs to the

http://dx.doi.org/10.1016/j.fcr.2016.07.001 0378-4290/© 2016 Elsevier B.V. All rights reserved.

Please cite this article in press as: Wang, J.-Y., et al., Exploring micro-field water-harvesting farming system in dryland wheat (Triticum aestivum L.): An innovative management for semiarid Kenya. Field Crops Res. (2016), http://dx.doi.org/10.1016/j.fcr.2016.07.001

G Model FIELD-6745; No. of Pages 12 2

ARTICLE IN PRESS J.-Y. Wang et al. / Field Crops Research xxx (2016) xxx–xxx

areas likely to suffer the most from global warming, and arid and semiarid areas account for 82% of its total land area. As in most other areas of Africa, agricultural system of Kenya is mainly featured by the scattered small holder farmers’ production pattern (Conway and Toenniessen, 2003). In absence of capital aid, it is impossible for small farmers to adopt the costly and complicated technologies such as drip irrigation technology (Rochecouste et al., 2015). Traditionally, the cropping pattern of semiarid Kenya is dominated by flat planting technology. Either average yield or water use efficiency (WUE) of dryland wheat in Kenya is not more than half the world average. Mean annual rainfall amount is about 700 mm in semiarid Kenya, but the rainfall is highly variable and mainly occurs in rainy seasons. Most areas of semiarid Kenya are located at the east Africa highland being adjacent to the equator, and there are relatively abundant rainfall and heat resources (Kaggwa et al., 2011). Innovative agricultural system, particularly low-cost but high-yield farming system is urgently needed to introduce to this area. Maximal utilization of rainwater resource has become a critical way to increase agricultural productivity and profitability in rain-fed agricultural areas. Over last decades, micro-field rain-harvesting farming system (MRFS) has been established extensively used in northwest China and some areas of South Korea, which significantly contributed to the improvements on field productivity of staple crops such as dryland wheat (Gan et al., 2013; Arnhold et al., 2013; Zhao et al., 2014). For the design principle of MRFS, ridges and furrows are established to harvest rainwater and store rainwater. The covered ridges serve as rainwater harvesting zones, and the furrows as planting zones. Meanwhile, the planting zones can collect and store rainwater from the runoff on the ridges. This system can collect water from light rain with almost 5 mm amount as minimum, and of course can retain surface runoff from heavy rain. Moreover, the system can reduce soil and water loss, and enhance carbon assimilation rate of atmospheric CO2 in drought-prone areas (Zhao et al., 2012; Arnhold et al., 2013). Therefore, the MRFS proves to be an innovative, high-yield and low-cost farming system, mainly including ridge-furrow mulching system and cropping system (Tian et al., 2003; Gan et al., 2013; Arnhold et al., 2013). The mulching materials generally include plastic sheets (black or transparent films), gravel sand, crop/grass straw and others, and the farming system is technically designed to optimize the proportion and size of ridge and furrow, the density of planting and the spatial layout of sowing (Gan et al., 2013). Since 1980s, the MRFS as a dominant tillage system was extensively used in semiarid northwest of China (Tian et al., 2003; Jia et al., 2006; Liu et al., 2009). Average grain yield and WUE of dryland wheat was up to 5374.5 kg ha−1 and 15.3 kg ha−1 mm−1 respectively, as a product of adopting the MRFS in northwest China (Liu et al., 2013; Zhang et al., 2013; Zhang et al., 2015). At field scale, the MRFS displayed great potential in improving field productivity and profitability (Wiyo et al., 1999; Wang et al., 2015), as well as soil and water conservation (Lentz and Bjorneberg, 2003; Stagnari et al., 2014). On the other hand, on-field rain-harvesting tillage technologies can be strategically modified and optimized to meet specific agro-climatic conditions. Straw mulching can conserve soil water and increase organic residues at the topsoil layer for better soil nutrient balance. Plastic sheet mulching acts as an efficient manner to prevent water loss and regulate soil temperature (Liu et al., 2011). No matter what sorts of materials used, mulching treatment plays a critical role on soil temperature & moisture, soil aeration and microbial diversity in soil system, and thus promotes mineralization process and nutrient release (Hankin et al., 1982; Li et al., 2004, 1999; Liu et al., 2012; Wang et al., 2014). Specifically, ridge and furrow with plastic mulching has been verified to be one of the most efficient measures to improve crop yield under the condition of low economic input (Tian et al., 2003; Liu et al., 2009; Zhao et al., 2012).

Due to the advantages of low cost and simple operation, the MRFS is easily adopted by smallholder farmers and extensively used in a large area. Considering local agro-climatic and soil conditions, we speculate that it could result in the increases in both wheat production and economic income in semiarid Kenya. From the perspectives of agricultural system innovation, we postulate that the MRFS would bring about the desired efficiency and profitability of dryland wheat production in the rainfed agricultural areas of Kenya and provide a new pathway to upgrade local farming system. We chose the experimental farm of KARI Katumani Research Center as study site, because of its typical representative of rainfed agro-ecosystem in Africa. To identify the respective effects of different mulching patterns, three major materials were used across two growing seasons, including black plastic sheet, transparent plastic sheet and grass straw. Soil water storage, water use efficiency, field productivity and economic returns under different planting patterns were evaluated and compared. The objectives of this study are: 1) to compare the differences in the productivity and water use efficiency of dryland wheat under MRFS for the first time; 2) to evaluate the dynamics of soil water storage along with soil profile; 3) to quantitatively evaluate the economic profitability of various treatments; and 4) to validate the postulation of this study on the basis of observation data. 2. Materials and methods 2.1. Description of study site A two-year field experiment was conducted in 2012 and 2013 at Katumani Centre, Kenya Agricultural Research Institute (KARI) (1◦ 35 S, 37◦ 14 E; altitude 1600 m). The study site represents typical semiarid continental monsoon climate type of East Africa Highland, with mean annual rainfall of about 700 mm and annual reference crop evaporation of about 1800 mm. The precipitation pattern in this region is of bimodal characteristics, with first rainy season from mid-March to late June, and second rainy season from mid-October to late December. The average maximum and minimum temperature were 24.7 ◦ C and 13.7 ◦ C respectively across two growing seasons. The fertility of surface soil is generally low. Soil texture is acidic sandy loam and has been classified as chromic luvisols, with a poor water holding capacity. The average bulk density is 1.54 g cm−3 (Table 1). Total precipitation within the growing season were 89.9 mm in 2012 and 138.3 mm in 2013 growing seasons. More importantly, it was noted that pre-sowing rainfall amount (within two months before sowing) was 390.4 mm and 84.5 mm in 2012 and 2013 growing seasons, respectively. To some extent, higher pre-sowing rainfall resulted in better soil water availability in 2012 growing season. Moreover, most of in-season rainfall was distributed at jointing and flowering stages, i.e. sensitive stages of water demand for wheat growth in 2012 growing season. While in 2013 growing season, over half of rainfall happened at maturity stage, which was negative to grain formation of wheat. 2.2. Experimental design and field management A schematic diagram of various treatments is presented in Fig. 1. Five treatments were designed as: 1) ridge and furrow totally covered by transparent plastic sheet (RFT), 2) ridge and furrow totally covered by black plastic sheet (RFB), 3) ridge and furrow totally covered by grass straw (RFS), 4) ridge and furrow without mulching (RF), and 5) traditional flat planting without mulching (CK). Both transparent and black plastic sheets are polyethylene with the thickness of 0.008 mm and the width of 120 cm (made by Lanzhou Gold Field Corporation of China, Lanzhou, China), while black film

Please cite this article in press as: Wang, J.-Y., et al., Exploring micro-field water-harvesting farming system in dryland wheat (Triticum aestivum L.): An innovative management for semiarid Kenya. Field Crops Res. (2016), http://dx.doi.org/10.1016/j.fcr.2016.07.001

G Model

ARTICLE IN PRESS

FIELD-6745; No. of Pages 12

J.-Y. Wang et al. / Field Crops Research xxx (2016) xxx–xxx

3

Table 1 Physical and chemical properties of soil in experimental field (0–60 cm depth) in KARI Katumani Centre, Kenya. Soil depth (cm)

0–20 20–40 40–60

SOC (g kg−1 )

23.80 20.74 18.23

AP (g kg−1 )

0.14 0.05 0.02

TN (g kg−1 )

0.91 0.81 0.79

pH

5.8 5.9 6.0

EC (us cm−1 )

59.50 54.75 51.85

TS (g kg−1 )

0.35 0.31 0.31

Bulk density (g cm−3 )

1.56 1.56 1.50

Soil particle size

Soil Texture

Clay (%)

Silt (%)

Sand (%)

0.85 3.02 3.05

19.28 30.63 23.66

79.88 66.36 73.32

Sandy loam Sandy loam Sandy loam

Note: SOC, Soil organic carbon; AP, available phosphorus; TN, total Nitrogen; TS, total salt content.

in the field. The grass straw was obtained from surrounding meadows before sowing, and then cut by hand into 5–10 cm length and air-dried. In RFS plots, the dried grass straw was applied manually on the same day at the rate of 6 t ha−1 in both growing seasons. All plots were weed-free and no fertilizer was applied for all treatments in both growth seasons based on the local farming practice. Plants were harvested on 22 August in 2012 and on 15 February in 2013, respectively. For the management of plastic residue, we had actually collected them by hands following the end of each growing season. The lands were remained clean to avoid any other influence on soil and water conservation in the experimental site across two growing seasons. 2.3. Sampling and measurements 2.3.1. Soil water content Soil water content (SWC, %) was determined gravimetrically for each two weeks at each 20-cm increment within the depth of 100 cm across the whole growing season. In each plot, soil samples were taken in the center of furrows with three replicates by using a soil auger (5 cm diameter, 20 cm height). The SWC was also measured before sowing and after harvesting. In the meantime, soil bulk density was determined throughout the soil profile (0–100 cm depth) and the average value is 1.49 g cm−3 . Soil water storage (SWS, mm) was calculated as follows: SWS = SWC × ␳b × H Fig. 1. Diagram of planting methods for various treatments in this study. (a) flat planting without mulching (CK); (b) ridge and furrow without mulching (RF); (c) ridge and furrow with transparent plastic mulching (RFT), with black plastic mulching (RFB) and with grass straw mulching (RFS).

can more efficiently suppress weeds growth and promote soil temperature compared with transparent one. The plastic sheets were laid out by hand over the plots where two pieces of plastic sheets were jointed in the middle of the furrow, and the joint was fixed stably by placing soil on the surface of the sheet. The ridge-furrow system was built with a manual machine in early May 2012 and early November 2012, respectively in two growing seasons. Prior to the experiment, the original state of study site was meadowland and no crops were planted. After the end of year 2012 trial, there was a fallow period of about 100 days till the start of year 2013 trial. Each treatment plot was replicated for three times and laid out in a randomized complete block design. Each plot was 5 m × 5 m = 25 m2 . Bare ridges were made around every plot to prevent runoff. In the meantime, a series of holes had been dug using iron wire at the film surface with an interval of 30 cm in each furrow. This doing help rainwater penetrate into the soil without runoff occurrence. The local wheat cultivar DUMA was planted on 17 May 2012 and 25 November 2012 using a manual hole-sowing machine in the furrow area at a seeding density of 450 plants m−2 with 20 cm row spacing. In all experimental plots, plant density was maintained as 450 plants m−2 , corresponding to local common plant density in the region. For the plastic sheet mulching plot, 0.29 kg plot−1 (equal to 116 kg ha−1 ) of plastic film was used

Where SWC is soil water content (%), ␳b is soil bulk density (g cm−3 ) and H refers to the thickness of the soil layer (mm). 2.3.2. Water use efficiency Water use efficiency (WUE) was calculated as the ratio of grain yield or above-ground biomass per unit area to total water consumption (evapotranspiration, ET) over the whole growing season. In study site, crop growth was completely dependent on precipitation and the precipitation was too low to bring drainage at below 1 m underground. There was no runoff due to the ridges around each ridge-furrow plot, and no irrigation was applied throughout the whole growing season. Seasonal evapotranspiration (ET) for each plot was determined using the equation: ET = P + W Where P is total precipitation in one growing season (mm) and W is the difference in SWS between before sowing and after harvesting. Water use efficiency of yield (WUEY ) and of above-ground biomass (WUEB ) was calculated as follows: WUEY = Y/ET WUEB = B/ET Where Y is grain yield (kg ha−1 ), B is above-ground biomass (kg ha−1 ) and ET is evapotranspiration amount in each growing season.

Please cite this article in press as: Wang, J.-Y., et al., Exploring micro-field water-harvesting farming system in dryland wheat (Triticum aestivum L.): An innovative management for semiarid Kenya. Field Crops Res. (2016), http://dx.doi.org/10.1016/j.fcr.2016.07.001

G Model FIELD-6745; No. of Pages 12

ARTICLE IN PRESS J.-Y. Wang et al. / Field Crops Research xxx (2016) xxx–xxx

4

2.3.3. Growth and yield components Growth traits were recorded in each fifteen days from 20 days after sowing (DAS) to maturity stage. Ten individual plants were randomly chosen in each plot and marked to be measured for plant height and leaf area. The leaf area was calculated as follows: Leaf area (cm2 ) = leaf length (cm) × leaf width (cm) × 0.83(Xiong et al., 2006) Meanwhile, the youngest fully expanded leaf was selected in each plant for measurement of the specific leaf area (SLA). The equation of the SLA is as follows: SLA(m2 /kg) = leaf area/leaf dry weight At harvesting stage, plants were sampled in 1.0 m2 area for determination on yield and yield components. Grain yield, aboveground biomass, spike number per m2 and seed number per spike were recorded for each plot. All the samples for biomass were put into the forced-air oven at 105 ◦ C for 1 h and at 80 ◦ C for a minimum of 72 h. 2.3.4. Cost and benefits analysis For the purpose of calculating labor cost, local farmers were employed to undertake the labor work. Values of labor input and materials investment are two major items for input sources in this study. In each treatment, the labor work included soil preparation, field management, sowing and harvesting. In RFT, RFB, RFS, and RF treatments, additional labor work also comprised ridge-furrow making, film/grass straw mulching and others. Referring to local wage level with 4.2 USD per farmer per workday (8 h), the labor cost of all employed farmers was converted into capital input in US dollars. In addition, the cost on materials in our study included plastic sheets and commercial seeds. There was no fertilizer application in all treatments. Weeds was controlled using hand weeding multiple times in the RFS, RF and CK treatments. For weed control, we only considered the labor cost in the input fund of labor (shown as field management in Table 6). Economic output was evaluated based on economic harvest of grain yield and hay yield. Considering local market at Machakos, the average price from 2012 to 2013 was 0.55 USD per kg for seed grain, and 0.03 USD per kg for hay yield respectively. Finally, net income in each treatment was determined by calculating the differences in the values between total output and total input.

2.4. Statistical methods Analysis of variance (ANOVA) was performed using the SPSS 17.0 program for Windows. To focus on identifying the effects of different rainwater harvesting farming technologies, the SWS data was harvested in different soil layers over two growing seasons and further subjected to an integrated analysis. Principal Component Analysis (PCA) and Person Correlation Analysis were also performed to determine the linkage among the parameters. T-test analysis was used for examining the differences in critical parameter(s) among five treatments across growing seasons. Figures were drawn using the Origin 8.0 software (OriginLab, USA, http://www. originlab.com/). Comparisons were conducted using least significant difference (LSD) at 0.05 probability level. Mean values are reported in the tables and figures. 3. Results 3.1. Variation of soil water storage (SWS) under the MRFS Rainwater-harvesting farming system led to significant improvement in soil water availability (0–100 cm) throughout whole growing season (Fig. 2). The SWS varied with soil depths and wheat growth stages (Figs. 2 and 3), and two growing seasons followed a similar trend in the dynamics of total SWS along with soil profile. Over whole growth period, total SWS showed a decreasing trend in all treatments, except for a temporal increase in ridge-furrow mulching treatments (RFT, RFB and RFS) at the initial stage of 2012 growing season (30 DAS). The base value of SWC was around 220 mm at the beginning of 2013 growing season, slightly higher than about 210 mm in 2012. The lowering rate of SWS in 2012 appeared to be less than that of 2013 in all the treatments, suggesting 2013 growing season was drier due to higher evaporation strength (as presented in the below). Critically, on-field rain-harvesting cropping treatments (RFT, RFB and RFS) led to better performance of SWS in the soil depths in two growing seasons. At the maturity stage, the RFT and RFB had the highest SWS, i.e. 180 mm and 150 mm in 2012 and 2013 growing seasons, respectively (Fig. 2). Generally, the dynamics of SWS varied from soil depths and growth stages in both growing seasons (Fig. 3). In 2012, the SWS in 0–20 cm soil layer was higher in mulching treatments than that in non-mulching treatments. At the 45th DAS, the SWS was increased to peak points in RFT and RFB treatments, with the percentage

Fig. 2. Time-course dynamics of total soil water storage in 0–100 cm soil profile in various treatments over two growing seasons. Bars stands for LSD at P ≤ 0.05. RFT, ridge and furrow with transparent plastic mulching; RFB, ridge and furrow with black plastic mulching; RFS, ridge and furrow with grass straw mulching; RF, ridge and furrow without mulching; and CK, traditional flat planting. The same below.

Please cite this article in press as: Wang, J.-Y., et al., Exploring micro-field water-harvesting farming system in dryland wheat (Triticum aestivum L.): An innovative management for semiarid Kenya. Field Crops Res. (2016), http://dx.doi.org/10.1016/j.fcr.2016.07.001

G Model FIELD-6745; No. of Pages 12

ARTICLE IN PRESS J.-Y. Wang et al. / Field Crops Research xxx (2016) xxx–xxx

5

Fig. 3. Time-course dynamics of soil water storage in different soil layers in various treatments over two growing seasons. Bars stands for LSD at P ≤ 0.05.

of increase by 184% and 189%, compared with CK, respectively (Fig. 3a). The SWS in subsoil (21–60 cm) was truly greater in three mulching treatments (RFT, RFB and RFS) than that of FR and CK. Among three mulching treatments, the SWS of RFT and RFB was significantly higher than that of RFS, except for at 15 DAS (Fig. 3b). In the 61–100 cm soil layer, there was no significant difference in SWS among the treatments at early growth stage. Afterwards, RFT and RFB gradually demonstrated obvious advantage in water storage, compared with RF and CK (P ≤ 0.05) (Fig. 3c). What’s more, the dynamics of SWS remained more stable in deep soil layer (61–100 cm) during the whole period of wheat growth in 2012. On the other hand, worse water storage was observed in 2013. The dynamic of SWS in all treatments showed similar trends (Fig. 3d–f). At early growth stage, there was no significant difference in SWS among the treatments in 21–100 cm soil layer, but a significant difference in topsoil (0–20 cm) among the treatments (Fig. 3d–f). In conclusion, in all the soil depths, the difference in SWS before sowing and harvesting (W) was truly lesser in 2012 growing season than that of 2013 growing season. The W in the mulching treatments were significantly smaller than that of RF and CK treatments in both growing seasons (Fig. 3 and Table 3).

3.2. Plant height and leaf area Both plant height and leaf area are the typical physiological indicators of plant growth, reflecting the dynamics of soil water availability. In general, both plant height and leaf area were significantly worse in CK than those of MRFS treatments, and two growing seasons shared similar time-course trends in both growth parameters in the treatments (Fig. 4). At early growth stage, there was no significant difference in plant height among the mulching treatments (RFT, RFB and RFS). Over time, plant height in either RFT or RFB treatment was increasingly greater than that of RFS treatment. Also, the values of plant height in RFT, RFB and RFS were becoming greater than those of RF and CK from the 20th DAS to maturity stage. Interestingly, there was no significant difference in plant height between RF and CK treatments across whole growth stage, except at the 80th DAS and at harvest in 2012. RFT treatment had the greatest plant height throughout growth stage, while there was no significant difference in comparison with RFB (P > 0.05). In general, plant height was truly significantly greater in 2012 growing season than that of 2013 growing season (P ≤ 0.05), and ridge-furrow tillage with plastic mulching led to the greatest plant height.

Please cite this article in press as: Wang, J.-Y., et al., Exploring micro-field water-harvesting farming system in dryland wheat (Triticum aestivum L.): An innovative management for semiarid Kenya. Field Crops Res. (2016), http://dx.doi.org/10.1016/j.fcr.2016.07.001

G Model FIELD-6745; No. of Pages 12

ARTICLE IN PRESS J.-Y. Wang et al. / Field Crops Research xxx (2016) xxx–xxx

6

Fig. 4. Time-course changes of plant height and leaf area in various treatments in two growing seasons. Bars stands for LSD at P ≤ 0.05.

Leaf area per plant tended to increase during vegetative growth stage and reached up to the maximal value since the beginning of reproductive stage. It turned to decrease after 50–60 DAS in all treatments for each growing season (Fig. 4). Prior to 20 DAS, there was no significant difference in leaf area among the treatments (P > 0.05). Afterwards, leaf area in RFT and RFB started to be greater than that of other treatments, and furthermore leaf area of RFS was significantly greater than that of RF and CK treatments. In 2012, the leaf area in MRFS treatments were increased by 34–58% in comparison with CK. In 2013, the leaf area in MRFS treatments were increased by 50–79% in comparison with CK. The ranking of leaf area in all treatments is as follows: RFT > RFB > RFS > RF > CK. 3.3. Yield components On-field rainwater-harvesting cultivation technologies significantly affected the yield composition of wheat in two growing seasons (Table 2). For spike number per unit area, there was no significant difference between RFT and RFB treatments. The spike number per unit area in CK was the least, and those of RF and RFS were in the middle. No significant difference was observed between RFS and RF treatments (P > 0.05). In 2012, the spike length in MRFS treatments were increased by 31–44%, compared with that of CK. Similarly, the spike weight in the MRFS treatments were increased by 16–35%, in comparison with that of CK (Table 2). However, there was no significant difference in gain weight per spike among mulching treatments, and there was no significant difference in RF and CK treatments. In 2013, the spike length in MRFS treatments were increased by 18–30%, compared with CK. However, there was no significant difference between RFT and RFB, RF and

CK. the spike weight in the MRFS treatments were 40–56% greater than that of CK. The grain number per spike in RFS treatment was significantly lower than that of RFT and RFB treatments, and there was a significant difference between RF and CK (P ≤ 0.05) (Table 2). Unlike other yield-related components, 1000-kernel weight was in general not affected by planting patterns according to our observation. An exception was that in 2012, 1000-kernel weight in RFT and RFB treatments was significantly lower than that of other treatments (P > 0.05), and there was no significant difference among the RFS, RF and CK treatments. In 2013, no significant difference was found among all five treatments (P ≥ 0.05). To say, the values of 1000-kernel weight were generally lower in ridge-furrow mulching treatments in comparison with CK, while the grain number per spike and the spike number per unit area were increased significantly (P ≤ 0.05) as a result of ridge-furrow mulching operation. 3.4. ET, grain yield, above-ground biomass and water use efficiency There were significant differences in evapotranspiration (ET, mm) among treatments over two growing seasons (Table 4). In general, ridge-furrow mulching led to better effect of evaporation suppression. Higher evaporation amount was observed in 2013 growing season (up to 220 mm), in comparison with about 120 mm in 2012 growing season. Both RFT and RFB treatments had the lowest ET among five treatments, and there was no significant difference between RFS and RFB treatments in two growing seasons (Table 4). Among the RFS, RF and CK, there was no significant difference in ET. Furthermore, CK kept the highest ET among all the treatments, up to 170.3 mm and 304.9 mm respectively in 2012 and

Please cite this article in press as: Wang, J.-Y., et al., Exploring micro-field water-harvesting farming system in dryland wheat (Triticum aestivum L.): An innovative management for semiarid Kenya. Field Crops Res. (2016), http://dx.doi.org/10.1016/j.fcr.2016.07.001

G Model

ARTICLE IN PRESS

FIELD-6745; No. of Pages 12

J.-Y. Wang et al. / Field Crops Research xxx (2016) xxx–xxx

7

Table 2 Yield-related components of wheat cv DUMA under different treatments in two growing seasons in KARI Katumani Centre, Kenya. Growing season

Field managements

Spike length (cm)

Gain number per spike

Gain weight per spike (g)

Spike weight (g)

1000-kernel weight (g)

Spike number per m2

First growing season

RFT RFB RFS RF CK

10.4 ± 0.18 c 10.1 ± 0.18 c 9.4 ± 0.27 b 7.6 ± 0.18 a 7.2 ± 0.32 a

48.6 ± 1.67 c 46.6 ± 0.76 c 40.6 ± 1.19 b 35.8 ± 0.84 a 34.5 ± 0.75 a

1.57 ± 0.08 b 1.56 ± 0.03 b 1.51 ± 0.02 b 1.31 ± 0.03 a 1.29 ± 0.03 a

2.66 ± 0.11 c 2.73 ± 0.15 c 2.34 ± 0.06 b 2.07 ± 0.03 a 2.02 ± 0.03 a

32.29 ± 1.0 a 33.50 ± 0.61 a 37.42 ± 1.01 b 36.60 ± 0.51 b 37.52 ± 1.07 b

406.4 ± 5.45 c 394.7 ± 2.07 c 358.5 ± 7.91 b 351.1 ± 2.70 b 259.5 ± 0.97 a

Second growing season

RFT RFB RFS RF CK

8.6 ± 0.20 c 8.3 ± 0.18b c 7.8 ± 0.29 b 6.8 ± 0.21 a 6.6 ± 0.23 a

37.5 ± 1.05 d 37.3 ± 0.56 d 29.5 ± 0.73 c 20.6 ± 0.76 b 18.0 ± 0.42 a

1.15 ± 0.03 d 1.18 ± 0.01 d 0.99 ± 0.02 c 0.66 ± 0.02 b 0.59 ± 0.01 a

2.03 ± 0.05 c 2.11 ± 0.03 c 1.89 ± 0.05 b 1.46 ± 0.03 a 1.35 ± 0.04 a

30.84 ± 1.35 a 31.62 ± 0.67 a 33.77 ± 0.78 a 32.44 ± 1.35 a 33.17 ± 1.11 a

376.0 ± 1.43 c 370.6 ± 1.62 c 348.2 ± 1.43 b 351.7 ± 4.77 b 295.6 ± 1.49 a

Notes: Different letters in the same growing season within a column stand for statistically significant differences at P ≤ 0.05. Values are showed as means ± S.E. of the mean. RFT, ridge and furrow with transparent plastic mulching; RFB, ridge and furrow with black plastic mulching; RFS, ridge and furrow with grass straw mulching; RF, ridge and furrow without mulching; and CK, traditional flat planting. The same below.

Table 3 Person correlation coefficients between yield and the yield components of wheat cv DUMA in KARI Katumani Centre, Kenya. Growing season

LA

LA

PL

SLA

TN

EL

GN

GW

EW

SN

EN

AB

HI

TKW

2012

PL SLA TN EL GN GW EW SN EN AB HI TKW Y

0.91** −0.31* 0.53** 0.80** 0.75** 0.60** 0.70** 0.43** 0.74** 0.71** 0.70** −0.45** 0.76**

−0.32* 0.62** 0.85** 0.82** 0.65** 0.69** 0.47** 0.89** 0.84** 0.77** −0.50** 0.88**

−0.37** −0.22 −0.30* −0.06 −0.18 0.02 −0.47 −0.32 −0.41 0.41** −0.37

0.51** 0.47** 0.28* 0.42** 0.23 0.56* 0.41 0.77** −0.38** 0.55*

0.84** 0.68** 0.71** 0.59** 0.73** 0.67** 0.79** −0.54** 0.76**

0.8** 0.79** 0.38** 0.72** 0.71** 0.72** −0.62** 0.77**

0.77** 0.36** 0.55* 0.62* 0.44 −0.03 0.62*

0.45** 0.60* 0.62* 0.68** −0.30* 0.68**

0.27 0.34 0.49 −0.17 0.41

0.94** 0.65** −0.32 0.93**

0.66** −0.19 0.98**

−0.53* 0.80**

−0.29

PL SLA TN EL GN GW EW SN EN AB HI TKW Y

0.91** −0.36* 0.63** 0.70** 0.91** 0.92** 0.87** 0.72** 0.75** 0.81** 0.80** −0.17 0.86**

−0.37** 0.66** 0.72** 0.92** 0.92** 0.87** 0.72** 0.79** 0.89** 0.77** −0.22 0.89**

−0.30* −0.37** −0.38** −0.42** −0.39** −0.33* −0.49 −0.56* −0.58* 0.025 −0.59*

0.64** 0.70** 0.64** 0.61** 0.65** 0.58* 0.62* 0.67** −0.26 0.68**

0.76** 0.74** 0.64** 0.72** 0.57* 0.66** 0.58* −0.25 0.67**

0.93** 0.88** 0.70** 0.81** 0.90** 0.86** −0.38** 0.93**

0.93** 0.76** 0.81** 0.87** 0.80** −0.03 0.89**

0.70** 0.80** 0.83** 0.72** −0.07 0.83**

0.55* 0.55* 0.58* −0.01 0.60*

0.96** 0.84** −0.10 0.94**

0.86** −0.21 0.98**

−0.23 0.94**

−0.23

2013

Notes: * P ≤ 0.05; **P ≤ 0.01. LA, leaf area; PL, plant height; SLA, specific leaf area; TN, tiller number; EL, ear length; GN, grain number; GW, grain weight; EW, ear weight; SN, spikelet number; EN, ear number per m2 ; AB, above-ground biomass; HI, harvest index; TKW, 1000-kernel weight; Y, yield.

2013 growing season. Comparatively, 2013 was much drier than 2012 growing season in terms of within-season aridity (ET/rainfall) in this study. As for grain yield per unit area, the 2012 growing season had better harvest performance than 2013 growing season did, and the RFT and RFB treatments showed the greatest yields among all the treatments (Table 4). Comparatively, the yield in CK was the lowest among all treatment in both seasons, i.e. 2039.7 kg ha−1 in 2012 and 1505.9 kg ha−1 in 2013, respectively. The yield performance of either RFS or RF was worse than that of RFT and RFB, but better than that of CK. We also calculated the percentage of increase in the yield among various treatments. In 2012, the yield in RFT, RFB, RFS and RF treatments were significantly increased by 60–130%, compared with CK, whereas there was no differences between RFS and RF. In 2013, the yield in RFT, RFB, RFS and RF treatments was significantly increased by 83–163%, compared with CK. Totally, ridge-furrow plastic mulching treatments had the best performance to improve

grain yield per unit area. Additionally, the dynamics of aboveground biomass showed similar trends among the treatments in both growing seasons. Furthermore, the above-ground biomass showed a similar trend with the yield in all treatments. Obviously, ridge-furrow mulching system enhanced the accumulation of total biomass in two growing seasons (Table 4). Water use efficiency (WUE) is a critical agronomic parameter to assess the effects of MRFS on field productivity (Table 4). We used two sorts of WUE indicators, including the WUEY as defined as the ratio of grain yield to total water consumption, and the WUEB as the ratio of biomass to total water consumption. Both WUE indicators followed similar trends in various treatments, whereas 2012 growing season generally had greater WUE than 2013 growing season did. For WUEY , the highest value in two growing seasons were found in the RFT treatment (Table 4). However, in 2012, no significant difference was observed in WUEY between RFT and RFB treatments (P > 0.05) (Table 4). Yet, when it came to 2013 growing

Please cite this article in press as: Wang, J.-Y., et al., Exploring micro-field water-harvesting farming system in dryland wheat (Triticum aestivum L.): An innovative management for semiarid Kenya. Field Crops Res. (2016), http://dx.doi.org/10.1016/j.fcr.2016.07.001

G Model

ARTICLE IN PRESS

FIELD-6745; No. of Pages 12

J.-Y. Wang et al. / Field Crops Research xxx (2016) xxx–xxx

Notes: Different letters in the same growing season within a column stand for statistically significant differences at P ≤ 0.05. Values are showed as means ± S.E. of the mean.

0.33 ± 0.01c 0.33 ± 0.01c 0.29 ± 0.01b 0.30 ± 0.01b 0.26 ± 0.01a 187 167 105 77 55.2 ± 1.1 d 51.3 ± 1.6 d 39.4 ± 1.1 c 34.0 ± 0.2 b 19.2 ± 0.2 a 3965.5 ± 154.1 c 3793.5 ± 199.1 c 3045.4 ± 21.5 b 2749.9 ± 120.8 b 1505.9 ± 8.4 a 216.4 ± 1.6 a 221.6 ± 5.6 a 262.8 ± 2.1 b 273.1 ± 1.8 b 304.9 ± 3.6 c 138.3 RFT RFB RFS RF CK 2013

78.1 83.3 124.5 134.8 166.6

163 152 102 83

18.3 ± 0.6 d 17.1 ± 0.5 d 11.6 ± 0.1 c 10.1 ± 0.4 b 4.9 ± 0.1 a

271 246 135 104

11938 ± 246 c 11361 ± 286 c 10358 ± 246 b 9282 ± 105 a 5863 ± 133 a

104 94 77 58

0.36 ± 0.01b 0.36 ± 0.01b 0.32 ± 0.01a 0.31 ± 0.01a 0.32 ± 0.01a 176 169 99 71 103.4 ± 1.9 c 100.4 ± 1.1 c 74.4 ± 9.3 b 64.1 ± 1.6 b 37.4 ± 1.3 a 103 89 84 63 4692.3 ± 201.5 d 4330.0 ± 183.8 cd 3817.7 ± 793.2 bc 3257.3 ± 296.7 b 2039.7 ± 194.8 a 125.0 ± 0.2 a 119.8 ± 0.3 a 158.9 ± 9.8 b 161.9 ± 1.5 b 170.3 ± 3.4 b 89.9 RFT RFB RFS RF CK 2012

35.1 29.9 69 72 80.4

130 112 87 60

37.6 ± 0.9 d 36.2 ± 0.8 c 23.9 ± 1.5 b 20.1 ± 0.9 a 12.0 ± 0.5 a

214 203 100 68

12916 ± 235 c 12023 ± 103 bc 11701 ± 152 bc 10382 ± 336 b 6365 ± 241 a

Increase in WIEB (%) WUEB (kg ha−1 mm−1 ) Increase in above-ground biomass (%) Above-ground Biomass (kg ha−1 ) WUEY Increase in (kg ha−1 mm−1 ) WUEY (%) Increase in Yield (%) Yield (kg ha−1 ) ET (mm) Rainfall W (mm) (mm) field managements Years

Table 4 Evapotranspiration, yield, biomass and water use efficiency of wheat cv DUMA in different treatments in two growing seasons in KARI Katumani Centre, Kenya.

Harvest Index

8

Fig. 5. PCA analysis on growth, yield and yield parameters in all treatments over two growing seasons. Yield, Y; leaf area, LA; plant height, PH; specific leaf area, SLA; tiller number, TN; ear length, EL; grain number, GN; grain weight, GW; ear weight, EW; spikelet number, SN; ear number per m2 , EN; above-ground biomass. AB; harvest index, HI; and 1000-kernel weight, TKW.

season, there was a significant difference between RFT and RFB. In 2012, the WUEY in RFT, RFB, RFS and RF treatments was significantly increased by 68–214%, compared with CK, while there was no difference between RF and CK. In 2013, the WUEY in RFT, RFB, RFS and RF treatments were significantly increased by 104–271%, compared with CK, and there was no significant difference between RFT and RFB. On the other hand, the WUEB showed a similar trend as WUEY in two growing seasons. As a result of ridge-furrow mulching, the WUEB ranged from 64.1 to 103.4 kg ha−1 mm−1 in 2012 and from 34.0 to 55.2 kg ha−1 mm−1 in 2013. Compared with CK, the WUEB in ridge-furrow plastic mulching treatments were increased by 76.8%–187.0% in both growing seasons (Table 4). Additionally, ridge-furrow mulching treatments (RFT and RFB) significantly increased the harvest index (HI), up to 0.36 in 2012 and 0.33 in 2013 respectively (Table 4) (P < 0.05). There was no significant difference in HI between RFT and RFB over two growing seasons. The HI values of other treatments ranged from 0.31 to 0.32 in 2012 and from 0.26 to 0.30 in 2013, respectively. There was no significant difference in HI between RFS and RF (Table 4). It can be concluded that ridge-furrow mulching treatments tended to rise up the reproductive allocation as well as total biomass production. 3.5. Linkage analyses and PCA between yield and its related components Correlations between yield and yield-related components were calculated among various treatments in each growing season (Table 3, Fig. 5). In 2012, either leaf area per plant or plant height was significantly positively associated with yield-related components (P < 0.01), but significantly negatively associated with specific leaf area (P < 0.05) and 1000-kernel weight (P < 0.01). In addition, specific leaf area had a significant positive association with 1000kernel weight and a negative association with tiller number and grain number per spike. Specific leaf area was of no significant relationship with other yield-related components. 1000-kernel weight showed a negative correlation with other yield components, and a positive correlation with specific leaf area. For the linkage analysis on yield formation, there were no significant correlations between grain yield and specific leaf area, spikelet number and 1000-kernel

Please cite this article in press as: Wang, J.-Y., et al., Exploring micro-field water-harvesting farming system in dryland wheat (Triticum aestivum L.): An innovative management for semiarid Kenya. Field Crops Res. (2016), http://dx.doi.org/10.1016/j.fcr.2016.07.001

G Model

ARTICLE IN PRESS

FIELD-6745; No. of Pages 12

J.-Y. Wang et al. / Field Crops Research xxx (2016) xxx–xxx

9

Table 5 Component score coefficient of yield components of PCA on wheat cv DUMA in KARI Katumani Centre, Kenya. Componentcategories

1 2

Component Score Coefficient of the yield formation factors LA

PL

SLA

TN

EL

GN

GW

EW

SN

EN

AB

HI

TKW

0.96 0.06

0.96 0.10

−0.03 0.86

0.83 −0.06

0.94 −0.06

0.95 −0.08

0.94 0.21

0.92 0.07

0.72 −0.04

0.60 −0.51

0.88 0.23

0.87 −0.26

0.19 0.80

Eigenvalues

Variance%

Cumulative%

8.49 1.84

65.34 14.13

65.34 79.47

Notes: LA, leaf area; PL, plant height; SLA, specific leaf area; TN, tiller number; EL, ear length; GN, grain number; GW, grain weight; EW, ear weight; SN, spikelet number; EN, ear number per m2 ; AB, above-ground biomass; HI, harvest index; TKW, 1000-kernel weight.

weight respectively (P > 0.05). Instead, the yield was significantly positively correlated with plant height (r = 0.88, P < 0.01), ear number per m2 (r = 0.93, P < 0.01) and above-ground biomass (r = 0.98, P <0.01). In 2013, there was a significant negative association between grain yield and specific leaf area (r = −0.59, P < 0.05), and a significant positive association between grain yield and spikelet number respectively (r = 0.60, P <0.05). No significant relationship was found between grain yield and 1000- kernel weight. As shown in Table 5, the relationship among the yield components showed similar trends in both growing seasons. More importantly, Fig. 5 and Table 4 indicate the results of principal component analysis (PCA) considering various yield components. For each variable considered, the eigenvector is represented. The proportion of variation explains 60.43% by the first axis and 17.82% by the second axis. The first axis is related mainly to morphological traits and ear organ traits, especially weight and number, while the second axis mainly represents specific leaf area, 1000- kernel weight and above-ground biomass (Fig. 5). 3.6. Profitability analysis Table 6 presents the results of economic analysis on the effects of MRFS. The wages of employed farmers were statistically combined into total economic input in this study. According to our working records, total expense of labor input in both growing seasons in all five treatments were from167.7 to 297.5 USD ha−1 . The labor input in RFT, RFB, RFS and RF treatments was increased by 20–113%, compared with CK. At average, the input of harvesting management was the same among all treatment. The expense of labor input in RFT, RFB, RFS and RF was higher than CK due to additional labor input in these treatments, including making ridges and furrows, purchasing plastic sheets, and mulching work. Combined with the input of material expense in each treatment, the sum of two growing seasons input in all five treatments were from 575.4 to 1224.6USD ha−1 . In comparison with CK, the net income in the ridge-furrow mulching treatments (RFT, RFB, RFS and RF) was significantly increased by 82–145%. We also calculated the ratio of output to input. Obviously, the lowest ratio was found in CK treatment, while the highest ratios were in the RFS treatment. On the basis of above analysis, it can be concluded that the ridge-furrow mulching system led to better economic benefits than that CK treatment (Table 6). 4. Discussion The innovation on agricultural system has received increasing concern worldwide under climate change and food shortage (Lehmann et al., 2013; Rochecouste et al., 2015). Improving rainwater utilization in dryland crops such as wheat is a key scientific issue in semiarid rainfed agricultural system. Our study first presented the agronomic performance and environmental adaptability of MRFS using local dryland wheat cultivar in a typical semiarid rainfed area of Kenya. The results showed that the MRFS achieved

the desired productive and economic outcomes in this area, leading to significant improvement on grain yield per unit area, soil water storage, dry-matter accumulation, WUE and economic benefits for dryland wheat. The outcomes have experimentally validated the reliability and feasibility of our hypothesis. To some extent, the MRFS appears to have greater productive potential and wider applicable scope in semiarid Kenya than semiarid China. It is noted that both growing seasons were defined as in dry years according to multi-year average precipitation in study site. Actually, local mean annual rainfall is around 700 mm, almost 2-fold that of semiarid China. Local annual reference crop evaporation is about 1800 mm and the aridity index (AI, the ratio of evaporation to precipitation) is around 2.57 (Mungai et al., 2000). However, mean annual evaporation and rainfall amounts are 1500 mm and 382 mm respectively, and the aridity index is 3.93 in semiarid China (Arora 2002; Zhao et al., 2014). Thus, the MRFS has greater productive potential and applicable perspective in semiarid Kenya even other large areas of semiarid Africa. In this study, the ridge-furrow mulching treatments were observed to produce outstanding effects on harvesting and utilizing rainwater, improving growth (leaf area and plant height etc.) and yield formation, and further achieving better economic benefits. Particularly, the ridge and furrow with plastic mulching obtained better performance than that of grass straw mulching. What’s more, the transparent plastic mulching was the best among the treatments. Our previous studies showed that transparent plastic mulching can increase topsoil temperature for better growth under the condition of insufficient accumulative soil temperature at early stage of cool season (Liu et al., 2009). Temporal distribution pattern of rainfall, together with MRFS technologies, played integrated influences on plant growth, yield formation and farmland hydrological process (Liu et al., 2012). It is needed to address the differences in seasonal rainfall amount and its time-course distribution pattern between two growing seasons. The rainfall amount in 2012 growing season was only 89.9 mm, and 2013 growing season’s rainfall amount was up to 138.3 mm. However, most of rainfall was accumulated at the early growth stage and the filling stage in 2012, which was well matched with sensitive stages of crop water demand. In contrast, most of the rainfall in 2013 took place at the maturity stage, i.e. missing the sensitive stage of crop water demand. Besides, the ET and aridity of 2012 growing season was lower than that of 2013 growing season. To sum up, temporal distribution characteristics of rainfall accounted for why higher grain yield, biomass and SWS were observed in 2012 but not in 2013 growing season (Fig. 2 and 3 and Table 4). On the other hand, hydrothermal mechanisms of MRFS improving plant growth and water conservation are discussed in the perspective of plant-soil interaction. The enhancement of crop production can be achieved primarily by capturing all available soil water by the end of life cycle (Rybka and Nita, 2015). Greater leaf area resulted from photosynthetic products accumulation under better soil moisture. During vegetative period, leaf area was quickly increased up to a status of canopy closure. Canopy shades turned to lower topsoil temperature and accordingly prevent water evapora-

Please cite this article in press as: Wang, J.-Y., et al., Exploring micro-field water-harvesting farming system in dryland wheat (Triticum aestivum L.): An innovative management for semiarid Kenya. Field Crops Res. (2016), http://dx.doi.org/10.1016/j.fcr.2016.07.001

G Model

ARTICLE IN PRESS

FIELD-6745; No. of Pages 12

4.28 4.06 5.86 5.34 3.86 5241.7 4972.8 4303.6 3685.2 2218.3 Notes: Price per unit for grain, 0.55 USD kg−1 . Price per unit for dry straw, 0.03 USD kg−1 (the price of seed grain and hay yield were cited from local market).

4755.9 4515.0 3847.7 3275.5 1957.8 485.9 457.8 455.9 409.7 260.5 1224.6 1224.6 735.0 690.6 575.4 240 240 240 240 240 389.6 389.6 0 0 0 RFT RFB RFS RF CK Sum of two growing seasons

108.4 108.4 108.4 108.4 108.4

243.3 243.3 193.3 171.1 113.5 RFT RFB RFS RF CK 2013

486.6 486.6 386.6 342.2 227

1802 1748 1600 1335 679 3.94 3.86 5.35 4.87 3.36 2414.3 2360.5 1967.3 1679.9 966.7 2175.1 2133.5 1748.0 1484.0 836.0 239.2 227.0 219.4 196.0 130.7 612.3 612.3 367.5 345.3 287.7 120 120 120 120 120 194.8 194.8 0 0 0

4.62 4.27 6.36 5.81 4.35 2827.5 2612.3 2336.2 2005.3 1251.6 2580.8 2381.5 2099.7 1791.5 1121.8 246.7 230.8 236.5 213.7 129.8 612.3 612.3 367.5 345.3 287.7 120 120 120 120 120 194.8 194.8 0 0 0 54.2 54.2 54.2 54.2 54.2 243.3 243.3 193.3 171.1 113.5 RFT RFB RFS RF CK 2012

54.2 54.2 54.2 54.2 54.2

2215 2000 1969 1660 964

Net income Output/Input Sum of output Grain yield Hay yield

Output value Sum of input

Plastic sheets

Commercialseeds Input fund of materials

MRFS making, sowing, field managements

Harvest Input fund of labor Treatments Growing season

Table 6 Economic benefits analysis on different treatments in two growing seasons in KARI Katumani Centre, Kenya. (Unit: USD ha−1 ).

4017 3748 3569 2995 1643

J.-Y. Wang et al. / Field Crops Research xxx (2016) xxx–xxx

10

tion from the soil. During the thriving stage of reproductive growth, getting a better hydrothermal balance was important for plant growth under the conditions of soil surface mulching (Weldearegay et al., 2012; Vial et al., 2015). Plastic mulching can directly affect the microclimate around the plants by modifying the radiation budget of field surface and decreasing soil water loss (Moreno and Moreno, 2008). Straw or plastic mulching usually reduces soil evaporation and stabilizes soil temperature, hence increasing crop yield (Vial et al., 2015). To say, mulching on soil surface can increase the reflection of solar radiation in the day and decrease the radiation heat loss from soil at night, resulting in increased minimum temperature and decreased maximum temperature and accordingly reducing diurnal variation in topsoil temperature (Gill et al., 1996; Vial et al., 2015). Advantageous soil water status and less variable soil temperature generally favor crop growth and yield (Tolk et al., 1999), particularly in Kenya where the field soil with low available water-holding capacity (Table 1). In this case, the ridgefurrow mulching treatments presented better yield performance and water use efficiency mainly due to optimal hydrothermal balance than the ridge-furrow non-mulching treatment did. With respect to the analysis on economic benefits, total input mainly comprised labor input and material expenses in this study. We also investigated local market and obtain the information of commercial prices of seed grain and hay. Generally, local labor input included the expenses of making ridge & furrow, sowing, field management and harvesting work. The material expenses mainly came from the procurements of plastic sheets (in MRFS) and commercial seeds. The economic input of harvesting work and purchasing commercial seeds were the same among all treatments (Table 6). Although the labor cost was higher in RFT, RFB and RFS treatments due to ridge-furrow making and plastic sheet/straw mulching (Table 6), the extent of economic harvest improvement in these treatments was generally greater than that of CK. For example, the sum of labor expense ranged from 450.6 USD ha−1 to 595 USD ha−1 in RFT system, only being increased by 47.6%–77.4% in comparison with that of CK (335.4 USD ha−1 ) in two growing seasons. Totally, the sum of net income in two growing seasons varied from 3568.6 to 4017.1 USD ha−1 in ridge-furrow mulching treatments, while only 1642.9 USD ha−1 in CK. The sum of net income in ridge-furrow mulching treatments was increased by 117%–145%, compared with CK. The magnitude of the increase in net income was higher than additional labor cost in two growing seasons. The ratio of output to input in RFS treatment was up to 5.86: 1, a highest value among all treatments in two growing seasons. This was mainly because grass straw was available for free, without additional expense as purchasing commercial plastic sheets (194.8 USD per hectare). Also, straw mulching required lower labor input than plastic film mulching did, according to our operating records of employed farmers. Therefore, total economic input was lowered in RFS treatment. In addition, straw mulching treatment had similar effects as plastic mulching treatment did on reducing soil moisture loss, collecting rainwater, and producing higher yield and economic output. However, the magnitude of SWS improvement in RFS was lower than that of plastic mulching treatments. That’s why the net income in RFS was lower than that of either RFT or RFB treatment (Table 6). In fact, plastic film is the most widely used due to its excellent properties and low cost (Gan et al., 2013; Hu et al., 2014). Facing with increasing population and climate change pressures on arable land, traditional flat planting pattern is hard to meet the demands of socio-economic development and food security in rainfed agricultural areas (Lehmann et al., 2013). Insufficient food production is a key issue affecting resident livelihood and ecosystem sustainability in African areas. It is reported that the rate of annual population growth is up to 3%, being much higher than the world average rate. Moreover, over half of the African populations live in rural areas and depend directly or indirectly on agriculture from the imme-

Please cite this article in press as: Wang, J.-Y., et al., Exploring micro-field water-harvesting farming system in dryland wheat (Triticum aestivum L.): An innovative management for semiarid Kenya. Field Crops Res. (2016), http://dx.doi.org/10.1016/j.fcr.2016.07.001

G Model FIELD-6745; No. of Pages 12

ARTICLE IN PRESS J.-Y. Wang et al. / Field Crops Research xxx (2016) xxx–xxx

diate environment for their livelihoods (Kiers et al., 2008). Wheat as a staple crop plays a critical role in ensuring food security and coping with climate change in Kenya. As a whole, Water-harvesting farming system in dryland wheat provides an innovation option for upgrading current agricultural system in semiarid Kenya and other Africa countries. Finally, this research is only focused on the specific site, and future efforts should be made for further experimental explorations in a larger area of Africa. 5. Conclusions Climate change has and will continue to alter the agro-climatic conditions for crop growth in semiarid rainfed agricultural areas of Africa such as semiarid Kenya. Due to historic and socioeconomic reasons, local rainwater resource use and field productivity was remained at low level in most areas of semiarid Africa including Kenya. It requires adjustments in management practices at the field scale. This study investigated a potential farming system to optimize rainwater collection and storage into soil system and lead to significant improvement on yield formation and field productivity in dryland wheat. It suggested that the MRFS played a critical role in significantly improving leaf area, ear length, grain number per spike and other yield components, which led to a great increase in grain yield per unit area in dryland wheat. Above-aground biomass and water use efficiencies were accordingly elevated in semiarid Kenya. More importantly, the farming practice of MRFS can produce higher economic ratio of output to input. The MRFS proved to be a cheap but efficient farming system to solve the issues of food security and agroecosystem sustainability. This study provides an innovation option to boost the productivity and profitability of dryland wheat in semiarid Kenya, and a potential solution to upgrade rainfed agricultural system in semiarid Kenya and even drying Africa. Acknowledgements We thank Drs. Wamuongo, Mureithi and Mukisira for collaborative assistance. The research was funded by International Cooperation Program of Ministry of Science and Technology of China (2015DFG31840 and 2014DFG32090), State Technology Support Program(2015BAD22B04), Fundamental Research Funds for the Central Universities of China (lzujbky-2015-br02) and Overseas Masters Program of Ministry of Education (Ms2011LZDX059). Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.fcr.2016.07.001. References Arnhold, S., Ruidisch, M., Bartsch, S., Shope, C.L., Huwe, B., 2013. Simulation of runoff patterns and soil erosion on mountainous farmland with and without plastic-covered ridge-furrow cultivation in south Korea. Trans. ASABE 56, 667–679. Arora, V.K., 2002. The use of the aridity index to assess climate change effect on annual runoff. J. Hydrol. 265, 164–177. Beddington, J.R., Asaduzzaman, M., Clark, M.E., Bremauntz, A.F., Guillou, M.D., Howlett, D.J.B., Jahn, M.M., 2012. What next for agriculture after Durban. Science 335 (6066), 289–290. Conway, G., Toenniessen, G., 2003. Science for African food security. Science 299, 1187. Gan, Y.T., Siddique, K.H.M., Turner, N.C., Li, X.G., Niu, J.Y., Chao, Y., Liu, L.P., Chai, Q., 2013. Ridge-furrow mulching systems-an innovative technique for boosting crop productivity in semiarid rain-fed environments. Adv. Agron. 118, 429–476. Garnett, T., Appleby, M.C., Balmford, A., Bateman, I.J., Benton, T.G., Bloomer, P.B., Burlingame, B., Dawkins, B., Dolan, L., Fraser, D., Herrero, M., Hoffmann, I., Smith, P., Thornton, P.K., Toulmin, C., Vermeulen, S.J., Godfray, H.C.J., 2013.

11

Sustainable intensification in agriculture: premises and policies. Science 341, 33–34. Gill, K.S., Gajri, P.R., Chaudhary, M.R., Singh, B., 1996. Tillage: mulch and irrigation effects on corn (Zea mays L.) in relation to evaporative demand. Soil Tillage Res. 39, 213–227. Hankin, L., Hill, D.E., Stephens, G.R., 1982. Effect of mulches on bacterial populations and enzyme activity in soil and vegetable yields. Plant Soil 64, 193–201. Hu, Q., Pan, F.F., Pan, X.B., Zhang, D., Yang, N., Pan, Z.H., Zhao, P.Y., Tuo, D.B., 2014. Effects of a ridge-furrow micro-field rainwater-harvesting system on potato yield in a semi-arid region. Field Crop Res. 166, 92–101. Jia, Y., Li, F.M., Wang, X.L., Yang, S.M., 2006. Soil water and alfalfa yields as affected by alternating ridges and furrows in rainfall harvest in a semiarid environment. Field Crop Res. 97, 167–175. Kaggwa, M.N., David, M.L., Michael, S., 2011. Small tropical reservoirs and fish cage culture: a pilot study conducted in Machakos district, Kenya. Aquacult. Int. 19, 839–853. Kiers, E.T., Leakey, R.R., Izac, A.M., Heinemann, J.A., Rosenthal, E., Nathan, D., Jiggins, J., 2008. Agriculture at a crossroads. Science 320, 320. Klerkx, L., Aarts, N., Leeuwis, C., 2010. Adaptive management in agricultural innovation systems: the interactions between innovation networks and their environment. Agric. Syst. 103, 390–400. Lehmann, N., Finger, R., Klein, T., Calanca, P., Walter, A., 2013. Adapting crop management practices to climate change: modeling optimal solutions at the field scale. Agric. Syst. 117, 55–65. Lentz, R.D., Bjorneberg, D.L., 2003. Polyacrylamide and straw residue effects on irrigation furrow erosion and infiltration. J. Soil Water Conserv. 58, 312–318. Li, F.M., Guo, A.H., Wei, H., 1999. Effects of clear plastic film mulch on yield of spring wheat. Field Crop Res. 63, 79–86. Li, F.M., Song, Q.H., Jjemba, P.K., Shi, Y.C., 2004. Dynamics of soil microbial biomass C and soil fertility in cropland mulched with plastic film in a semiarid agro-ecosystem. Soil Biol. Biochem. 36, 1893–1902. Liu, C.A., Jin, S.L., Zhou, L.M., Jia, Y., Li, F.M., Xiong, Y.C., Li, X.G., 2009. Effects of plastic film mulch and tillage on maize productivity and soil parameters. Eur. J. Agron. 31, 241–249. Liu, Y., Shen, Y.F., Yang, S.J., Li, S.Q., Chen, F., 2011. Effect of mulch and irrigation practices on soil water soil temperature and the grain yield of maize (Zea mays L.) in Loess Plateau, China. Afr. J. Agric. Res. 6, 2175–2182. Liu, Y.J., Mao, L., He, X.H., Cheng, G., Ma, X.J., An, L.Z., Feng, H.Y., 2012. Rapid change of AM fungal community in a rain-fed wheat field with short-term plastic film mulching practice. Mycorrhiza 22, 31–39. Liu, Y., Han, J., Wen, X.X., Wu, W., Guo, Q., Zeng, A., Liao, Y.C., 2013. The effect of plastic-covered ridge and furrow planting on the grain filling and hormonal changes of winter wheat. J. Integr. Agric. 12, 1771–1782. Moreno, M.M., Moreno, A., 2008. Effect of different biodegradable and polyethylene mulches on soil properties and production in a tomato crop. Sci. Hortic. (Amst.) 116, 256–263. Morton, J.F., 2007. The impact of climate change on smallholder and subsistence agriculture. Proc. Natl. Acad. Sci. U. S. A. 104, 19680–19685. Mungai, D.N., Stigter, C.J., Coulson, C.L., 2000. Simply obtained global radiation, soil temperature and soilmoisture in an alley cropping system in semi-arid Kenya. Theor. Appl. Climatol. 65, 63–78. Rochecouste, J.F., Dargusch, P., Cameron, D., Smith, C., 2015. An analysis of the socio-economic factors influencing the adoption of conservation agriculture as a climate change mitigation activity in Australian dryland grain production. Agric. Syst. 135, 20–30. Rybka, K., Nita, Z., 2015. Physiological requirements for wheat ideotypes in response to drought threat. Acta Physiol. Plant. 37, 1–13. Stagnari, F., Galieni, A., Speca, S., Cafiero, G., Pisante, M., 2014. Effects of straw mulch on growth and yield of durum wheat during transition to conservation agriculture in mediterranean environment. Field Crop. Res. 167, 51–63. Tian, Y., Su, D.R., Li, F.M., Li, X.G., 2003. Effect of rainwater harvesting with ridge and furrow on yield of potato in semiarid areas. Field Crop Res. 84, 385–391. Tolk, J., Howell, T., Evett, S., 1999. Effect of mulch, irrigation and soil type on water use and yield of maize. Soil Tillage Res. 50, 137–147. Townes, O.C.H., 2015. Increase water harvesting in Africa. Nature 519, 283–285. Vial, L., Lefroy, R., Fukai, S., 2015. Application of mulch under reduced water input to increase yield and water productivity of sweet corn in a lowland rice system. Field Crop Res. 171, 120–129. Wang, Y.P., Li, X.G., Hai, L., Siddique, K.H.M., Gan, Y.T., Li, F.M., 2014. Film fully-mulched ridge-furrow cropping affects soil biochemical properties and maize nutrient uptake in a rainfed semi-arid environment. Soil Sci. Plant Nutr. 60, 486–498. Wang, Q., Ren, X., Song, X.Y., Hu, G.R., Zhang, E.H., Wang, H.L., Maureen, M.V., 2015. The optimum ridge-furrow ratio and suitable ridge-covering material in rainwater harvesting for oats production in semiarid regions of China. Field Crop Res. 172, 106–118. Weldearegay, D.F., Fang, Y., Jiang, D., Liu, F.L., 2012. Independent and combined effects of soil warming and drought stress during anthesis on seed set and grain yield in two spring wheat varieties. J. Agron. Crop Sci. 198, 245–253. Wiyo, K.A., Kasomekera, Z.M., Feyen, J., 1999. Variability in ridge and furrow size and shape and maize population density on small subsistence farms in Malawi. Soil Tillage Res. 51, 113–119. Xiong, Y.C., Li, F.M., Zhang, T., 2006. Performance of wheat crops with different chromosome ploidy: root-sourced signals, drought tolerance, and yield performance. Planta 224, 710–718.

Please cite this article in press as: Wang, J.-Y., et al., Exploring micro-field water-harvesting farming system in dryland wheat (Triticum aestivum L.): An innovative management for semiarid Kenya. Field Crops Res. (2016), http://dx.doi.org/10.1016/j.fcr.2016.07.001

G Model FIELD-6745; No. of Pages 12 12

ARTICLE IN PRESS J.-Y. Wang et al. / Field Crops Research xxx (2016) xxx–xxx

Zhang, S.L., Sadras, V., Chen, X.P., Zhang, F.S., 2013. Water use efficiency of dryland wheat in the Loess Plateau in response to soil and crop management. Field Crop Res. 151, 9–18. Zhang, P., Wei, T., Wang, H.X., Wang, M., Meng, X.P., Mou, S.W., Zhang, R., Jia, Z.K., Han, Q.F., 2015. Effects of straw mulch on soil water and winter wheat production in dryland farming. Nature 5, 1–12. Zhao, H., Xiong, Y.C., Li, F.M., Wang, R.Y., Qiang, S.C., Yao, T.F., Mo, F., 2012. Plastic film mulch for half growing-season maximized WUE and yield of potato via

moisture-temperature improvement in a semi-arid agroecosystem. Agric. Water Manag. 104, 68–78. Zhao, H., Wang, R.Y., Ma, B.L., Xiong, Y.C., Qiang, S.C., Wang, C.L., Liu, C.A., Li, F.M., 2014. Ridge-furrow with full plastic film mulching improves water use efficiency and tuber yields of potato in a semiarid rainfed ecosystem. Field Crop Res. 161, 137–148.

Please cite this article in press as: Wang, J.-Y., et al., Exploring micro-field water-harvesting farming system in dryland wheat (Triticum aestivum L.): An innovative management for semiarid Kenya. Field Crops Res. (2016), http://dx.doi.org/10.1016/j.fcr.2016.07.001