Case study 3: Water-saving rice agro-ecosystems for West Africa
Rice is the most consumed cereal in West-Africa and its demand is strongly increasing mostly due to population growth and changes in consumption pattern linked to increased urbanization (Elbehri, Kaminski, Koroma, Iafrate & Benali 2013; ECOWAS 2019). Currently, local production covers only roughly 60% of the demand, the rest of the populations’ needs being met through imports that strongly impact the region’s economy and make it vulnerable to changes in prices on the global commodity market (ECOWAS 2019). Since a majority of West Africans are projected to live in urban areas in the future, demand for rice is expected to increase further by 73.5 percent from 2011 to 2025 (Elbehri et al. 2013; ECOWAS 2019). Several programs have been launched to increase local production to meet this future demand and guarantee food sovereignty.
However, irrigated rice cultivation requires a large amount of freshwater. It was estimated that, in a dry environment with an evaporation rate higher than the precipitation rate found in large parts of West Africa, traditional irrigated rice cultivation requires between 700 to 1,500 mm of water to produce 1 kg of rice (Bhuiyan 1992). Furthermore, climate change is trending towards hotter and dryer atmosphere prevailing in the region, which will increase evaporative demand. With increased competition from industries and city growth for freshwater and increased uncertainty in precipitation patterns resulting in reduced water availability in the region, it will not be possible to meet the growing demand with local rice production using the conventional irrigated rice cultivation system (Nie et al. 2012). Several water-saving alternatives have been suggested including alternate wetting and drying (AWD) or aerobic cultivation in order to reduce water consumption and increase cultivated surfaces. While these practices can save up to 50% of the water used for rice production, they often incur a yield penalty with the current varieties that have been selected for irrigated agro-ecosystems (Bouman, Peng, Castañeda & Visperas 2005; Peng et al. 2006; Kreye et al.2009; Sasaki et al. 2010).
Aerobic rice cultivation aims to maximize crop water use efficiency by growing plants in soil without flooding or puddling (Matsunami, Matsunami & Kokubun 2009; Matsuo & Mochizuki 2009). It allows greater water savings and can be deployed in regions without access to irrigation water but has a high yield penalty and is associated with increased weed management and risks of nematodes. In this system, periodic drought stress may reduce yield stability and yield potential (Sandhu et al. 2019). For these reasons, aerobic rice has not become popular among farmers in irrigated areas (Meena, Bhusal, Kumar, Jain & Jain 2019). On the other hand, AWD is a simple practice where, instead of keeping the fields permanently flooded, irrigation is periodically stopped until the soil water table reaches a certain depth, easily measured using a pipe set in the soil, and then re-started until the field is flooded again (Bouman & Tuong 2001). AWD cycles are repeated either during the vegetative or flowering stage or throughout the rice cultivation cycle, although keeping paddies flooded in hot environments during the flowering stage help avoid the problems linked to heat sensitivity though the cooling effect of transpiration (Jagadish, Murty & Quick 2015). Two types of AWD have been described: moderate AWD when field water level is allowed to drop down to 15 cm below the soil surface, and severe AWD when soils are allowed to dry beyond −20 kPa (Carrijo et al. 2017). AWD can be easily adopted in these areas as it does not change the cultivation practices, is not associated with increased labour needs and can contribute to a reduction in water consumption of 5 to 30%, depending on the season and soil, as well as reduce methane emissions and grain arsenic levels (Linquistet al. 2015; Carrijo, Lundy & Linquist 2017).In general, AWD has a limited yield penalty (5.4% in a meta-analysis of 56 studies) but the yield decrease is more important in severe AWD or if AWD is maintained throughout the crop cycle (Carrijo et al. 2017). The impact on yield is also very dependent on the genotype and most of the currently used high-yield varieties show yield reduction in AWD (Carrijoet al. 2017; Sandhu et al. 2017). Thus, there is a clear need to develop new varieties to optimize yield in AWD rice agroecosystems.
The use of root traits has been little explored in such agroecosystems, but could have a large impact to optimize Water Use Efficiency (WUE), P use efficiency (PUE) and N use efficiency (NUE) and thus reducing inputs globally in AWD systems. Indeed, AWD results in periodic changes in water content in the topsoil but also changes the dynamics of nutrient availability and in particular of N and P availability (Wang et al. 2016; Acosta-Motos et al. 2020). For instance, AWD increases topsoil P availability and has been linked with changes in the soil microbiota and the stimulation of aerobic P-solubilizing bacteria in the aerobic topsoil compartment (Li et al.2018). Conversely, N availability seemingly decreases upon AWD due to increased denitrification, volatilisation and leaching although these losses can be avoided by timely N application (Tan et al. 2013; Djaman et al. 2018). Water and N signalling are known to interact in ways that affect root phenes for synergistic or antagonistic resource uptake (reviewed in Araus, Swift, Alvarez, Henry & Coruzzi 2020). Hence, varieties for AWD agro-ecosystems need to be adapted to fluctuations in soil water and nutrient content in the topsoil. Root phenes could be mobilized to that end as a large genetic diversity for these phenes is available in rice (Ahmadi et al. 2014). Interestingly, AWD seems to have a positive effect on root development as indicated by increased root biomass and maximum root length (Wanget al. 2016; Acosta-Motos et al. 2020).
Root phenes that increase the volume of topsoil exploration and promote P uptake should be prioritized to improve rice yields in AWD systems. Phenes that merit investigation include crown root number, lateral root density (there are two types of lateral roots in rice, thin and determinate short lateral roots and indeterminate long lateral roots; Rebouillatet al. 2009), and root hair length and density. A recent large-scale study by Sandhu et al. (2017) is consistent with nodal root number and lateral root density positively impacting yield in AWD. Evaluation of new rice varieties derived from crosses between popular varieties and drought-tolerant accessions was performed in fully-irrigated and AWD systems in seven sites across Asia (Sandhu et al. 2017). Out of 82 lines tested in at least three sites, lines with stable and high yield in AWD conditions compared to irrigated conditions were identified. Comparison of the root phenotype of a subset of these stable high-yielding lines and the control line IR64 (high yield variety that shows a reduction in yield in AWD) showed that higher nodal root number and root dry weight at 10–20 cm depth played an important role in maintaining grain yield under AWD (Sandhu et al. 2017). Interestingly, quantitative changes in these root phenes induced after initiation of AWD cycle, i.e. plasticity response of the root system, were shown to be important for yield stability (Sandhu et al. 2017). These observations are in agreement with previous analyses describing the important role of root system plasticity for drought and low P tolerance in field and controlled conditions (Sandhu et al. 2016). In the best performing lines under AWD (initially selected as drought tolerant breeding lines), the number of nodal roots below 20 cm and deep root length at the flowering stage appear to have a positive effect on yield (Sandhu et al. 2017). Therefore, these lines may have had the ability to grow roots in deeper soil layers for improved water and nitrogen acquisition in the dry down period (particularly at flowering stage in severe AWD scenario and less so when the water table is maintained above 15-30 cm), while AWD-induced root branching plasticity in the topsoil improved P uptake (Fig. 2C). Moreover, a study performed using a combination of pot experiments and functional-structural plant model indicated that, for P uptake, root system efficiency is improved by increased root branching both in irrigated and water stress conditions (De Bauw et al. 2020). In AWD, phosphate acquisition was shown to occur mainly at the root tip and led to phosphorus depletion along the root due to the low mobility of P in the soil (De Bauw et al. 2020). Root interactions with soil symbiotic microbes such as arbuscular mycorrhizal fungi (AMF), which is inhibited by flooding, might also improve water and nutrient acquisition in the topsoil (Vallino, Fiorilli & Bonfante 2014; Mbodj et al. 2018). Furthermore, AMF were shown to confer drought tolerance in rice (Chareesri, De Deyn, Sergeeva, Polthanee & Kuyper 2020). Interestingly, AMF colonization was also shown to reduce P loss from paddy fields thus improving PUE and reducing environmental impacts (Zhang et al. 2020). Thus, root phenes that promote AMF infection in topsoil such as increased number of crown roots and large lateral roots with more cortex cells and less aerenchyma to accommodate intracellular fungal structures may lead to improved water and nutrient uptake. Moreover, rice response to AMF infection is dependent on the plant genotype and could be targeted for plant breeding (Diedhiouet al. 2016; Huang et al. 2020; Lefebvre 2020).
Overall, improving root phenes in rice for AWD systems could lead to improved WUE but also PUE (Acosta-Motoset al. 2020) and NUE (Wang et al. 2016) through synergistic interactions between water and nutrients, thus leading to more sustainable rice agro-ecosystems with reduced water and fertilizer consumption. Lines adapted to aerobic conditions that show root plasticity in response to AWD in topsoil (increased branching) could be interesting materials to test. Rhizosphere phenes could also be targeted to improve water and nutrient acquisition efficiency in this system.