This study examined bat activity levels in rice fields, with the aim of exploring why rice fields set in different landscapes see different levels of insectivorous bat activity. Our results showed that key landscape features, namely forest and edge cover, urban land, moonlight intensity and landscape heterogeneity influenced the activity of five sonotypes of insectivorous bats.
In warm and dry environments, bats lose water rapidly (Webb et al. 1995) and it must be replenished through their food and by drinking. While many species have evolved mechanisms to limit water loss (Reher and Dausmann 2021), the presence of water bodies still exert considerable influence on where bats choose to fly, particularly in arid environments (Razgour et al. 2010). Past studies have found distance to water to be a key factor in driving selection of roost sites and foraging grounds (Adams and Thibault 2006; Adams and Hayes 2008; Rainho and Palmeirim 2011). However, water bodies in these studies tend to be few and far apart, increasing their importance to foraging bats. Despite our study being conducted in the summer, when temperatures regularly exceeded 30°C, having been conducted in rice fields, our sites were all within 600m of a water body (defined here as areas of water of at least 1000sq meters, equivalent to a 50m by 20m plot), with an average distance of 204m. This distance being well within the foraging ranges of insectivorous bats, water was not found to be an influencing factor in the best model of any sonotype, suggesting that for the analysed species, the range at which water becomes a limiting factor is greater than bats in our landscapes were presented with.
Urban landscapes are not entirely uninhabitable, and many bat species have adapted to them. The use of urban land by bats is influenced by built infrastructure, light pollution, noise levels, tree cover, bat physiology, predation pressure, and prey availability (Moretto and Francis 2017; Moretto et al. 2019; Jung and Threlfall 2021). While light pollution in urban areas is harmful to insect populations on the large scale (Owens et al. 2020), the attraction provided by street lights can create local prey-dense zones (Firebaugh and Haynes 2019), which in turn can increase the activity of some bats (Rodríguez-Aguilar et al. 2017). Bat activity is higher when patches of forest, water, and/or agricultural landcover are nearby (Dixon 2012; Ancillotto et al. 2019), in part because of higher insect activity in such areas (Avila-Flores and Fenton 2005). Studies have found that the complementation of anthropogenic and natural landcover can result in high levels of bat activity, particularly of mobile generalist species (Johnson et al. 2008). More universally, canopy cover has been found to be a key determinant of bat activity in urban contexts (Bailey et al. 2019). Bats exhibit extremely species-specific responses to landscapes and while some studies have found increased activity of bats in urban landscapes (Rodríguez-Aguilar et al. 2017), many studies have found even moderate urbanization to have negative impacts on bat activity (Ancillotto 2015; Jung and Threlfall 2016). While our sites were set in rice fields, urban landcover was present around the sites in two forms: dense urban landscapes seen in Nagaon city and scattered nodes in villages. One or both of these negatively affected the activity of S28 and S38. For the other sonotypes, it is possible that the farmland and associated buildings, trees, and water bodies satisfied the requirements of most bats at the local scale, thereby overriding any negative (or positive) effects of the city.
Forests and forest boundaries, including linear features such as rows of trees, are important habitats for bats (Heim et al. 2015). However, separating the often concurrent incentives for the use of such habitats, including as roosting sites, for safety against predators (Heim et al. 2018) and the wind (Verboom and Huitema 1997), landmarks for navigation, and improved foraging (Jantzen and Fenton 2013), can be a challenge. Some of these benefits change with time. For example, brighter nights allow predators to see better (Prugh and Golden 2014), increasing the importance of shelter. Similarly, prey abundance relative to the surrounding landscape varies with season, particularly in agricultural landscapes. Conversely, the provision of trees as roosting sites remains more consistent over time.
Prey availability is a strong driver of bat activity and while it is difficult to compare the activity of different bats in different settings, studies have found that insectivorous bats do, in general, consider in-season rice fields to be prey-rich habitat (Sedlock et al. 2019; Toffoli and Rughetti 2020). Activity is increased by the presence of forest patches nearby (Heim et al. 2015; Bailey et al. 2019). Some studies have shown insectivorous bats to preferentially hunt over rice fields compared to forested areas (Puig-Montserrat et al. 2015; Kemp et al. 2019; Katunzi et al. 2021). Other studies report that while rice fields were attractive foraging grounds, nearby natural wetlands saw higher levels of foraging activity (Toffoli and Rughetti 2017).
While the edge variables calculated and tested for in this study included area of forest edge (within various buffers from the focal site), only distance to the nearest edge was retained in the best models of S28 and S32. The activity of sonotypes S35 and S38 was influenced by the distance to the nearest forest patch while S28 was influenced by the area of forest within 1km. Given that our landscapes were surrounded by forested area, rather than natural wetlands, we hypothesize that our rice fields were more attractive foraging sites than the surrounding forests or forest edges and therefore that the primary attraction of forests and forest edges was not prey availability but safety or the presence of roost sites. This is augmented by the fact that the bamboo houses and sheds in villages were known to be used as roost sites (IB, rigorous personal observation). These structures, being smaller than concrete buildings, more enclosed by trees, and bearing thatched roofs, were usually classified as ‘edge’ rather than ‘urban’ landcover. This adds to the argument that ‘edges’, as classified in this study, also represented roost sites to bats in the study area.
Area of rice was not present in the best models of any sonotype. The most likely reason being that since our sites were all in rice fields, set in rice-dominated landscapes, prey availability was high enough that larger areas did not improve foraging success through reduced competition or increased prey availability.
Rice-dominated landscapes present a harsh matrix interspersed with patches of forest-urban-edge combinations that provide foraging grounds, roost sites, and relative safety. Harsher (more contrasting) matrices evoke stronger reactions to fragmented landscapes (Rodríguez-San Pedro and Simonetti 2015; Farneda et al. 2020) and unbroken rice fields provide little by way of shelter or roost sites. Greater landscape heterogeneity increases the proximity of different landcover types. This can result in larger bat populations and higher bat activity because bats that roost and forage in different landscapes now have easy access to both, reducing potentially commuting costs (Ethier and Fahrig 2011). An increase in landscape heterogeneity also correlates with availability of edges between natural and agricultural landcover. These interfaces are important for edge-space foragers that avoid open spaces either because they are more vulnerable to predators in open spaces, or because they have a higher foraging success at edges (Lentini et al. 2012; Frey-Ehrenbold et al. 2013). Greater heterogeneity also increases insect populations by providing spaces for them to breed and survive fallow seasons (Sigsgaard 2000; Fahrig et al. 2015; Bertrand et al. 2016; Chaperon et al. 2022). Sonotype S47 was influenced by landscape heterogeneity, most strongly at the scale of 1km (Fig. 2), supporting other studies that demonstrate the positive effect of landscape heterogeneity on bat activity (Frey-Ehrenbold et al. 2013; Monck-Whipp et al. 2018; Rodríguez-San Pedro et al. 2019).
Agricultural industries globally face an enormous challenge to reduce their reliance on environmentally unsustainable practices, such as the use of chemical pesticides, for both environmental and financial reasons. Pests are predicted to cause increasing losses to many major crops (Deutsch et al. 2018) and chemical methods of control, far from eliminating pests, have often prompted the emergence of more resistant strains (Normile 2013). The effective use of natural enemies as a sustainable alternative is only practised in pockets around the world (Lou et al. 2013; Puig-Montserrat et al. 2015), but holds great potential to alleviate the dual burdens of declining agricultural and ecological health. Realizing this potential requires both an understanding of the ecological relationships that underpin these ecosystem services and the ecology of the service providers to make agricultural landscapes more habitable to them.
Studies of bats in agricultural landscapes have shown the need for connectivity, heterogeneity, and the presence of natural landcover (Downs and Racey 2006; Frey-Ehrenbold et al. 2013; Finch et al. 2020). Few have considered the rice landscape, and none in India. This study demonstrated that in rice landscapes of Assam, while proximity to urban land decreased bat activity, the presence of forest patches and edges, and higher levels of landscape heterogeneity promoted bat activity, and likely by extension also increased the control of insect pests.