1 | INTRODUCTION
Aquatic invasions have become common and are one of the greatest threats to aquatic ecosystems worldwide ( Alexander et al., 2015; Gherardi, 2007; Sala et al., 2000) by altering nutrient cycling (Vanni, 2021), food webs (DeBoer et al., 2018), and abundance and distribution of native species (Gallardo et al., 2016; Weber & Brown, 2011). Currently, bigheaded carps (Hypothalmichthys moltrix andHypopthalmichthys nobilis ) are some of the most problematic aquatic invasive species that have spread rapidly and broadly in North America with adverse ecosystem effects (DeBoer et al., 2018; Solomon et al., 2016; Tillotson et al., 2022). Pelagic planktivores native to China and a small portion of eastern Russia (Kolar et al., 2007), bigheaded carp were introduced to the United States for aquaculture and sewage treatment resulting in their eventual escapement to natural waters where populations rapidly increased (Kolar et al., 2005). Since their initial escapement in the early 1980s, the Mississippi River has served as an invasion highway throughout the central United States. The Upper Mississippi River (UMR) is currently one of the primary invasion fronts for bigheaded carp in the Midwestern United States due to a series of locks and dams that have slowed their upstream expansion (Fritts et al., 2021; Tripp et al., 2013; Whitledge et al., 2019). Lock and Dam 19 specifically represents a major barrier to their northern expansion ( Larson et al., 2017; Fritts et al., 2021; Tripp et al., 2013), although adults have been captured as far north as pool 2 (U.S. Geological Survey, 2022).
While adult bigheaded carp have been captured up to pool 2, reproduction has only been documented to pool 16 (Camacho et al., 2023; Larson et al., 2017). Fish reproductive phenology is the product of many environmental conditions (Krabbenhoft et al., 2014; Yang et al. 2021) and quantifying fish reproduction provides critical temporal, spatial, population, and community metrics in relation to environmental conditions (Pritt et al., 2015; Quist et al. 2004). Seasonality in concert with water temperature is often considered necessary for successful reproduction by triggering the release of gametes (Pankhurst & Porter, 2003; Werner, 2002). Bigheaded carp require water temperatures > 17°C to initiate reproduction, with reproductive activity occurring up to 30°C (DeGrandchamp et al., 2007; Kocovsky et al., 2012; Schrank et al., 2001). Alternatively, reproduction of native fishes is generally initiated at cooler temperatures and occurs over a more narrow range (e.g., walleyeSander vitreus , 5 to 10°C; Bozek et al., 2011; gizzard shadDorosoma cepedianum , 10 to 21°C; Becker, 1983; freshwater drumAplodinotus grunniens , 18 to 25°C; Swedberg and Walburg, 1970). Beyond temperature, changing river discharge strongly affects reproductive activity in many invasive and native lotic fishes (Humphries et al., 2002; King et al., 2016; Dudley & Platania, 2007) including bigheaded carp (e.g., Kolar et al., 2007; Schrank et al., 2001).
Beyond environmental conditions, spatial variation in riverine habitats can influence where invasive and native fish reproduce (Camacho et al., 2023; Kolar et al., 2007). In rivers, spawning habitat relates to channel position such as channel border, thalweg, and backwaters due to differences in river discharge and species spawning requirements. Larval habitat use often corresponds to their reproductive guild (Holland, 1986). Some species such as bigheaded carp and freshwater drum spawn in the open water environments of the thalweg where eggs passively drift until they hatch and reach sizes that allow increased mobility to escape the current (Becker, 1983). Other species reproduce in the lower velocities in channel borders and backwaters that provide cover for eggs and developing larvae (e.g., walleye) or passively drift downstream (e.g., gizzard shad; Holland, 1986; Simon, 1998). Consequently, while most lotic larval fish assessments focus on thalweg collections, larval occupancy may vary among taxa and habitats.
Given the network of environmental conditions for fish reproduction, along with the uncertainty of ichthyoplankton sampling, determining occurrence of riverine fish reproductive events can be difficult. High spatio-temporal variability in larval catches is common (Cyr et al., 1992; Leonardsson et al., 2016; Michaletz & Gale, 1999; Weber et al., 2021). Therefore, it is difficult to determine when, where, and under what conditions fish reproduction occurs, particularly when detection probability is imperfect. Additionally, adult bigheaded carps are difficult to capture compared to adult native fishes (Bouska et al., 2017), but whether or not differences in sampling efficiency also exist at the larval phase is unknown. Occupancy (Ψ) and detection (p ) modeling offers a quantitative method of estimating the true presence of species in a system by accounting for imperfect detections based on discrete encounters over temporal or spatial scales (MacKenzie et al., 2002). Within fisheries, occupancy and detection modeling has been primarily used to describe distribution of adult fishes (Potoka et al., 2016; Schumann et al., 2020; Sullivan et al., 2018;), but models are plastic and have been adapted to evaluate age-0 fish habitat use (Burdick et al., 2008; Falke et al., 2010, 2012), timing of reproduction events associated with environmental conditions (Falke et al., 2010; Peoples and Frimpong, 2011; Pritt et al., 2014), and recruitment of cryptic invaders (Weber & Brown, 2019). These models also provide sample size requirements to achieve desired detection probabilities, improving sampling designs and monitoring programs (Kelly et al., 2021; Kuehne & Olden, 2016; Rodtka et al., 2015). Consequently, occupancy models can serve as a useful tool for assessing reproductive dynamics of bigheaded carp along leading edges of invasion where reproduction may be limited, and spawning events can be brief and difficult to detect.
Our objectives were to estimate occupancy and detection probability of bigheaded carp larvae and native fish taxa in relation to spatial and environmental variation. Next, we quantified cumulative detection probabilities to estimate sampling effort required to detect bigheaded carp larvae compared to native species. First, we hypothesized bigheaded carp larval occupancy would be higher in the thalweg compared to other habitats and positively related to water temperature and river discharge. Second, we hypothesized bigheaded carp larval occupancy would be lower than native fishes and lower in Mississippi River pools along the leading edge of population expansion than already established native fish populations. Finally, we hypothesized larval detection probabilities would be lower for bigheaded carp than native taxa.