Predator-prey Research Articles

Anthropogenic and environmental risk factors of salmonid predation in a tidal freshwater delta

Abstract Water diversions that support agricultural and municipal use result in fish mortality through entrainment and impingement. Additionally, this infrastructure may attract both predators and prey fishes, thereby increasing predation rates and prey mortality near these anthropogenic contact points. The Sacramento–San Joaquin Delta (the Delta) in California's Central Valley is a tidal freshwater ecosystem that exports large volumes of water for municipal and agricultural use while at the same time providing valuable migratory and rearing habitat for imperilled fishes. Emigrating juvenile salmonids experience high mortality in the Delta, with predation by non‐native fishes contributing substantially. Therefore, this study had three main objectives. First, we determined if small water diversions aggregated piscivorous fishes like other similar structures in freshwater ecosystems. Second, we determined how small diversions may influence juvenile salmon predation risk in conjunction with other known predation risk factors (e.g. predator abundance, temperature and depth). Third, we assessed the predator assemblage, abundance and distribution to determine the likely predator composition in objectives one and two. Throughout the spring of 2021, we used ARIS (adaptive resolution imaging sonar; Sound Metrics) sonars to compare piscivore abundance at 30 water diversions in the north Delta to shorelines adjacent to diversions that did not contain these structures. We used predation event recorders (PERs) to assess the predation risk juvenile salmonids were exposed to, with linear distance (m) from diversions, and other predation risk factors in the north Delta. Finally, we used a boat electrofishing survey to determine the piscivore assemblage and compare spatial trends in black bass (Micropterus spp.) CPUE and relative abundance throughout these waterways. Piscivore abundance was greater near small water diversions than at adjacent shorelines and the predation risk of juvenile salmonids increased with diversion proximity. Additionally, predation risk increased with increasing piscivore abundance and decreasing water depth. The north Delta predator assemblage was dominated by black basses (Micropterus spp.), which likely drove the negative relationship of predation risk with water depth, given habitat requirements of these species. Furthermore, increasing smallmouth (Micropterus dolomieu) and spotted bass (Micropterus punctulatus) abundance in our northern study sites may have weakened temperature effects on predation, given metabolic requirements of these species. Our work demonstrated that small water diversions are likely to increase mortality of endangered salmonids, and that the north Delta predator assemblage was different than recorded by previous work in this system, changing predation risk factors. Although more work is needed to determine the population level impacts of diversions, the ubiquitous distribution of these structures warrants management solutions to reduce mortality from this source. These results indicate that in addition to entrainment and impingement, water diversions may increase mortality of small‐bodied fishes by attracting predators and elevating predation risk. Given the continual human demand for freshwater, predator–prey interactions should be considered along with entrainment and impingement when assessing intake infrastructure mitigation, especially when diversions co‐occur along migratory routes and essential habitat of imperilled fishes.

Bifurcation analysis of a singular predator–prey plankton system with biotoxin delay and constant harvesting

Bifurcation analysis of a singular predator–prey plankton system with biotoxin delay and constant harvesting

Global bifurcation analysis and derivation of an optimal harvesting policy of a prey–predator model with Holling type-IV functional scheme

ABSTRACT This paper establishes a prey–predator model with a Holling type IV functional response and proportional harvesting in prey species and nonlinear harvesting in predator species. We provide a completely local and global bifurcation analysis of the model to characterize some meaningful results that may emerge from the interaction of biological resources. The proposed model exhibits various complex behaviors involving heteroclinic and homoclinic orbits, and we prove analytically that it also undergoes transcritical, Hopf, and Bogdanov–Takens bifurcations in the prey–predator plane. Moreover, the system is unfolded to compute the bifurcation curves, highlighting the critical dynamics by perturbing two bifurcation parameters near the cusp. Furthermore, from the economic point of view, the bionomic equilibrium of the system is studied, and optimal harvesting policy is derived using the Pontryagin’s maximum principle. Numerical simulations are presented to verify our analytical results.

Multiple Time Scales and Normal Form of Hopf Bifurcation in a Delayed Reaction–Diffusion–Advection System

In the reaction–diffusion systems, the Laplacian is usually a self-adjoint operator. Incorporating advection terms generally destroys this and yields a quite complicated decomposition of phase space in the Center Manifold Reduction (CMR) method. Considering that the Multiple Time Scales (MTS) method can be directly applied to bifurcation analysis and normal form derivation based on algebraic operations, and the computational process is relatively universal, we apply the MTS method to analyze the Hopf bifurcation problem of reaction–diffusion–advection systems with time delay. First, we introduce the MTS method for the system with a specific boundary condition, which is used to derive the normal form of Hopf bifurcation. Calculating the normal form using the MTS method can be summarized as determining the bifurcation parameters, conducting Taylor expansion based on the MTS idea, and eliminating secular terms. Then, the key coefficients in the normal form are explicitly given, which are also compared with the results obtained by the CMR method, and both methods lead to the same bifurcation results. Finally, we use the MTS method to calculate the normal form of Hopf bifurcation for a predator–prey model with mixed boundary conditions. We find that the spatially nonhomogeneous periodic solutions appear near the equilibrium in the system when the time delay exceeds the critical value, and the theoretical results are illustrated by numerical simulations.

Who is afraid of modelling time as a continuous variable?

Abstract Most models of ecological and eco‐evolutionary processes involve creating trajectories of something, be it population densities, average trait values, or environmental states, over time. This requires decision‐making regarding how to represent the flow of time in models. Most ecologists are exposed to continuous‐time models (typically in the form of ordinary differential equations) as part of their training, especially since the famous Lotka‐Volterra predator–prey dynamics are formulated this way. However, few appear sufficiently well trained to produce their own work with continuous‐time models and may lack exposure to the true versatility of available methods. Specifically, knowledge that discrete individuals can be modelled in continuous time using the Gillespie algorithm is not as widespread as it should be. I will illustrate the flexibility of continous‐time modelling methods such that researchers can make informed choices, and not resort to discretizing time as a ‘default’ without a clear biological motivation to do so. I provide three example‐based tutorials: (1) a comparison of deterministic and stochastic dynamics of the Lotka‐Volterra predator–prey model, (2) an evaluation of matelessness in a hypothetical insect population (and of selection to mate more often by either searching more efficiently or by shortening the ‘time out’ after each mating) and (3) within‐season density dependence followed by a birth pulse leading to Beverton‐Holt or Ricker dynamics depending on whether the deaths of conspecifics help reduce the mortality of others or not (compensatory mortality). I highlight properties of the exponential distribution that, while counter‐intuitive, are good to know when deriving expected lifetime reproductive success or other similar quantities. I also give guidance on how to proceed if the so‐called memorylessness assumption does not hold in a given situation, and show how continuous and discrete times can be freely mixed if the biological situation dictates this to be the preferred option. Continuous‐time models can also be empirically fitted to data, and I review briefly the insight this gives into the so‐called ‘do hares eat lynx?’ paradox that has been plaguing the interpretation of the Hudson Bay hare and lynx dataset.

Methylmercury Concentrations More Strongly Associated With Trait Variation Than Food Web Position in Plethodontid Salamanders.

Salamanders serve as bioindicators of mercury (Hg) in both terrestrial and aquatic habitats and are an important link in the food web between low-trophic prey and higher-trophic predators. We investigated the drivers of methylmercury (MeHg) exposure in three common plethodontid salamander species in New York State, USA, including comparisons among regions, habitat types (terrestrial and semiaquatic), and color morphs of Plethodon cinereus (striped and unstriped). Nonlethal tail samples were collected from one terrestrial species (P. cinereus) and two semiaquatic species (Eurycea bislineata and Desmognathus spp.) in the Adirondack Mountains (ADK) and the Finger Lakes National Forest (FLNF) regions. Samples were analyzed for MeHg and stable isotopes, including δ15N and δ13C which are proxies of trophic position and diet, respectively. Despite elevated biota Hg concentrations typically found in the ADK, salamander MeHg concentrations did not differ by region in the terrestrial species and one of the semiaquatic species. In addition, diet and trophic level did not explain MeHg exposure in salamanders. Semiaquatic salamanders exhibited higher MeHg concentrations than terrestrial salamanders in the FLNF only. Within species, only snout-vent length predicted MeHg concentrations in E. bislineata with few other variables significant as predictors of MeHg concentrations in path models. Among P. cinereus individuals in the FLNF, the striped morph had greater MeHg concentrations than the unstriped morph, and food web tracers were not different between morphs. Overall, New York State salamander Hg concentrations were elevated compared to other locations where these species are present. The present study establishes baseline Hg data in salamanders for future assessments of changes in Hg bioavailability to forests of New York State. Environ Toxicol Chem 2024;43:2045-2057. © 2024 SETAC.

Local and global dynamics of a prey–predator system with fear, Allee effect, and variable attack rate

Fear prompts prey to adopt risk-averse behaviors, such as reduced foraging activity, increased vigilance, and avoidance of areas with high predator presence, which affects its reproduction. In a real scenario, a population requires a minimum density to avoid extinction, known as an Allee threshold. In light of these biological factors, we propose a predator–prey model with (i) a fear effect in a prey population, (ii) an Allee effect in a predator population, and (iii) a non-constant attack rate that modifies the functional response. We ensured the non-negativity and boundedness of the solutions and examined the local and global stability status for each existing steady state solutions. We investigated some deep dynamical properties of the system by varying different parameters, such as cost of fear in prey and strength of the Allee effect in predators and their mortality rate. In codimension one bifurcations, we observed saddle node, Hopf, homoclinic, and coalescence of two limit cycles. Additionally, codimension two bifurcations were observed, including Bautin and Bogdanov Takens bifurcations. To provide a clearer understanding of these bifurcations, we conducted biparametric analysis involving the fear and Allee parameters, as well as the fear parameter and predator mortality rate. Our investigation shows that cost of fear and strength of Allee strongly influences the survival status of the predator. Furthermore, bistability and tristability reveal that the survival and extinction of predator are dependent on the initial population level. Numerical simulations and graphical illustrations are provided to support and validate our theoretical findings.

Mesopredators have differing influences on prey habitat use and diel activity in a multipredator landscape

AbstractResource distribution, habitat structure, and predators greatly influence spatial and temporal landscape use by prey species. The “risky places” hypothesis establishes prey will proactively respond to predators' presence based on habitat cues, whereas the “risky times” hypothesis predicts prey will reactively respond by increasing vigilance in the presence of predators regardless of habitat cues. We fit a multiscale, Bayesian species interaction occupancy model with detection/non‐detection data to evaluate black‐tailed jackrabbit (Lepus californicus) and eastern cottontail rabbit (Sylvilagus floridanus) habitat use in the presence and absence of coyotes (Canis latrans), American badgers (Taxidea taxus), and swift foxes (Vulpes velox). We also evaluated how species‐specific predator presence modified temporal activity patterns of prey. Jackrabbits decreased habitat use in areas with greater forage and opted to use areas with greater visibility when coyotes or swift foxes were present. However, cottontails used habitat in open areas with greater visibility when American badgers were present and all other predators absent, suggesting dissimilar habitat‐use patterns dictated by predator‐specific risks. Both lagomorph species are nocturnal with segregated peaks of activity compared with predators, suggesting fine‐scale temporal use partitioning. Our results provide insights into predator–prey dynamics across heterogenous landscapes in a multi‐predator system.

Thermodynamic Sustainability Assessment of Turkish Cash and Feed Crops: Evaluations and Recommendations

ABSTRACTThis work analyzes the sustainability status of cash crops (rice, wheat, and sunflower) and feed crops (silage and clover) cultivation in Turkey through thermodynamic sustainability assessment technique, emergy analysis (EA). EA offers insights into agricultural system's dependence on external resources, environmental loading, and renewability by classifying the resources that drive agricultural production as renewable, nonrenewable, and purchased. Data belonging to 17 agricultural systems evaluated in this work are gathered through survey method by interviewing the farmers. EA results reveal that production of cash crops is essentially unsustainable due to high dependence on externally purchased inputs fertilizer, diesel, and nonrenewable input water. Of cash crops, rice production has the lowest system renewability and the highest environmental loading. In feed crops, silage production is mainly unsustainable and clover production is in transition state in terms of sustainability. Feed crop production creates less environmental loading than cash crop production mainly due to being partially integrated with husbandry. Coinciding with this, the level of organic fertilization is found to be the determining factor in the sustainability status of the feed crop systems. Transforming these systems to sustainable ones requires increasing system's self‐sufficiency through enhanced circularity (increased nutrient recycling) and functioning of natural processes. Consequently, we recommend growing crops in polycultures rather than monocultures to benefit from sister plant and predator–prey relations, integrating crop production with animal rearing (natural fertilization) and utilizing waste‐origin bioenergy. For resource efficiency, we suggest adopting methods such as drip irrigation and utilization of biodegradable polymer‐coated fertilizer grains to prevent fertilizer runoffs.

Time-dependent probability density functions and information geometry in a stochastic prey–predator model of fusion plasmas

Time-dependent probability density functions and information geometry in a stochastic prey–predator model of fusion plasmas

Nonlinear dynamics in a fear‐driven predator–prey system: Bistability, bifurcations, hydra effect, and optimal harvesting

The impact of predator‐driven fear on ecosystems is significant and can encompass both trophic (direct) and nontrophic (indirect) effects. Previous studies have shown that nontrophic fear effects have an important role in predator–prey dynamics. This study investigates the nontrophic fear effect on prey caused by generalist predators and explores optimal harvesting. We assume that the reproduction rate of prey is reduced due to the fear effect, and generalist predator follows Holling type II foraging strategy for predation. Additionally, we assume that predators are commercially valuable and harvested proportionately to their density. We demonstrate the existence of equilibrium points, their local and global stability, and bifurcation analysis. We observe that the model system undergoes a sequence of codimension one and codimension two bifurcations. Our results show that in the absence of predator harvesting, increasing levels of fear destabilize the predator–prey system, and controlled harvesting is beneficial for the coexistence of both populations. Also, the harvesting of predators may produce hydra and multiple hydra effects. We identify different types of bistability phenomena, which emphasize the importance of initial population size. Further, an optimal harvesting policy is also explored by formulating an optimal control problem (OCP). The harvesting cost‐functional is designed by incorporating the bionomic equilibrium state. We use Pontryagin's maximum principle and solve the OCP numerically. It is observed that implementing optimal harvesting not only contributes to the ecological benefits by maintaining a sustainable balance of predator–prey evolution but also plays a significant role in maximizing the economic benefits.

A complete analysis of a slow–fast modified Leslie–Gower predator–prey system

In this paper, we transform a modified Leslie–Gower predator–prey system into the corresponding fast–slow version by assuming that prey reproduces much faster than predator, and then perform a complete analysis about its dynamics. More specifically, we find the necessary and sufficient conditions of the exact number (zero, one or two) of positive equilibria of the slow–fast system and its (or their) location, and then we further fully determine its (or their) dynamics under explicit conditions. Besides, by converting the slow–fast system into its slow–fast normal form, we are able to characterize its rich dynamics completely, including relaxation oscillation, singular Hopf bifurcation, canard explosion, homoclinic orbits, heteroclinic orbits and global attraction of equilibrium. Moreover, the sufficient conditions to guarantee these various rich dynamics are explicitly given, including the explicit conditions to determine whether singular Hopf bifurcation is supercritical or subcritical, which generally cannot be explicitly derived in the existing literatures. Additionally, the cyclicity of diverse canard cycles is found under explicit conditions. Of particular interest is that we show the existence and uniqueness of one canard cycle without head whose cyclicity is at most two under explicit parameters conditions. Our results complement and enrich the previous work (relaxation oscillation, and global attraction of a boundary equilibrium) about this fast–slow system. We also employ numerical simulations to illustrate our theoretical analysis.

Combination of a membrane bioreactor with a rotating biological contactor holding several diverse metazoans can reduce excess sludge with fouling mitigation

In a membrane bioreactor (MBR) system, in situ sludge reduction techniques induce membrane fouling. To address this challenge, we incorporated a rotating mesh carrier, which can adsorb organic matter and provide a habitat for metazoans, into the anoxic tank of a conventional anoxic/oxic-MBR (A/O-MBR) system, termed rotating biological contactor-MBR (RBC-MBR), and evaluated treatment performance. Over 151 days, lab-scale RBC-MBR and A/O-MBR were used to treat municipal sewage. Both reactors showed similar COD and NH4+ removal rates. However, RBC-MBR reduced excess sludge by approximately 45 % compared with A/O-MBR. Microscopic observation and 18S rRNA gene-based microbial analysis revealed the persistence of microfauna and metazoans (oligochaetes, nematodes, and rotifers) in RBC, which are typically absent in activated sludge. Additionally, the metazoan's population in the RBC-MBR membrane tank was two-fold that of A/O-MBR, indicating enhanced sludge reduction through predation. Despite these reductions, the increase in transmembrane pressure was similar between RBC-MBR and A/O-MBR, suggesting that sludge holding by RBC mesh media degrade fouling substances, such as proteins and polysaccharides and improves sludge filterability, resulting in membrane fouling mitigation. Microbial communities in both reactors were similar, indicating that the installation of RBC did not alter the microbial community of sludge. Network analysis suggested potential symbiotic or prey–predator relationships between bacteria and metazoans. This study reveals that RBC-MBR effectively reduced the excess sludge while mitigating membrane fouling, highlighting one of the promising technology for applying metazoan predation into MBR.

Feeding ontogeny: A new approach to predator–prey interactions

AbstractDiets of fish larvae may vary within populations and among species and are affected by larval prey availability and foraging capacity. We aimed to describe the diet of fish larvae based on empirical field observations, emphasizing the interaction and organization of the predator–prey network using Auchenipterus osteomystax as a species model. The network showed few interactions (low connectance and modularity) but a nested structure (some items that had fewer records tended to occur together with the most frequent ones), low complementary specialization (the proportions of ingested items tended to be equivalent) and a pattern of segregation (some items tended not to be consumed together) in the use of resources among individuals. Robustness was low when the removal of prey with a higher consumption frequency occurred. The correlation between morphological variables and network attributes was significant and positive for abundance and strength, while the correlation with nested rank was negative. The availability of suitable dietary resources is likely to be the key factor for success and survival in the early stages of development. With changes in morphology during larval development, A. osteomystax showed a generalist position in the network as its predation capacity increased. A new approach to trophic ecology studies of fish larvae incorporating network analysis may help explain the individual characteristics of larvae and assess the pattern of interactions in the population.

Stationary distribution and mean extinction time in a generalist prey–predator model driven by Lévy noises

In this study, a delayed generalist prey–predator model incorporating the logistic-like growth and Holling type-III functional response for predator subjected to two Lévy noises is established. Such environmental fluctuations can describe abrupt change in population density. First, we conduct stability analysis and determine the critical value of time delay when Hopf bifurcation takes place. Then, the existence of the global positive solution and stability in distribution of stochastic model are studied. Numerical results are given to illustrate the analytical findings, and mainly dedicated to the consequences of time delay and Lévy noises on the stationary distribution, mean extinction time and reproduction rate of prey–predator populations, via stationary probability distribution function (PDF), mean first passage time (MFPT) and relaxation time, respectively. Our findings indicate that the decreases in time delay, noise intensities (Di,i=1,2), and stability indexes within the range of αi∈[1,2] are more likely to maintain the population in a high density state. Phase transition occurs when the system is in the suitable parameter regime. Additionally, there exist optimal noise intensities that slow down switching of prey species from a stable state to an extinction state. Smaller time delay τ and stability indexes αi inhibit population extinction, whereas smaller skewness parameter β speed up prey extinction. More importantly, two noises can tune the enhanced stability characteristic of the system.

Evaluating the tipping point of a complex system: The case of disruptive technology

AbstractComplex systems often operate in equilibriums that can be disrupted under specific conditions, driving the system to undergo an irreversible phase transition (aka tipping point) into a new equilibrium. Disruptive technology, as introduced in the book Innovator's Dilemma, is a new emerging technology that can successfully displace incumbent technologies and push the market through phase transition into a new equilibrium. In this paper, the authors model the market disruption caused by a disruptive technology as a complex system, with dynamics that show a phase transition or tipping point, after which the system shifts into a new equilibrium aiming at assessing the success or failure of a disruptive technology. The purpose of this study is to mathematically model a tipping point measure of complex networks and a theoretical framework for disruptive technology dynamics. A predator–prey model is used to emulate the behavior of a disruptive technology versus an incumbent technology, and a resilience index is integrated to measure the tipping point in the market where the disruptive technology will overtake the incumbent technology. This methodology is applied to a historical case study of film versus digital versus cell phone cameras, demonstrating the application of this methodology. The study's contribution can be applied to both systems engineering and disruptive innovation management. The proposed approach can help stakeholders assess the impact of a potential new disruptive technology and use the resulting resilience index as a measurement for adjusting technology requirements and systems management approaches to achieve a desired outcome.

Hopf bifurcation analysis of a two‐delayed diffusive predator–prey model with spatial memory of prey

In this paper, we consider a diffusive predator–prey model with spatial memory of prey and gestation delay of predator. For the system without delays, we study the stability of the positive equilibrium in the case of diffusion and no diffusion, respectively. For the delayed model without diffusions, the existence of Hopf bifurcation is discussed. Further, we investigate the stability switches of the model with delays and diffusions when two delays change simultaneously by calculating the stability switching curves and obtain the existence of Hopf bifurcation. We also calculate the normal form of Hopf bifurcation to determine the direction of Hopf bifurcation and the stability of bifurcation periodic solutions. Finally, numerical simulations verify the theoretical results.

Evidence for a by‐product mutualism in a group hunter depends on prey movement state

Abstract Group‐hunting animals have been shown to engage their prey in long pursuits (kms) to increase capture success, but the evidence for this is limited to a few terrestrial species. This predation strategy may be effective in the open‐ocean, where group‐hunts are characterised by large predator–prey size ratios and there are few places for prey to hide. However, recording the attack rates, capture rates and capture success probabilities of grouping predators during long‐distance pursuits through the open ocean, remains a significant challenge. Here we use a combination of underwater and aerial video to assess the attack and capture behaviour of striped marlin, Kajikia audax, in relation to the movement of their schooling prey (‘stationary’ and ‘mobile’ movement states) during group hunts. We found that the attack rate was higher when the prey school was ‘stationary’ and, during these times, individual marlin could make longer sequences of uninterrupted attacks on the prey school, effectively monopolising the prey resource for short periods over conspecifics. When the prey school was ‘mobile’, the attack pattern changed and individual marlin alternated their attacks on the prey school more frequently. Capture success probability (likelihood of capture per attack) was not affected by prey movement state. However, attacks on ‘mobile’ schools resulted in more prey fish becoming isolated from the school, and these isolated prey were easily captured—not by the attacker but by conspecifics swimming nearby. The likelihood of a prey becoming isolated from a ‘mobile’ school was predicted by prey school speed, with a 0.1 m s−1 increase leading to an 18.01% increase in the chance of isolation. Altogether this meant the capture rate at the group level was highest when the prey school was ‘mobile’ rather than ‘stationary’ because the predators that were pursuing the ‘mobile’ prey school had an increased availability of isolated and easy to capture prey. This mechanism provides evidence for a by‐product mutualism selecting for conspecific group formation and we discuss our findings in consideration of the social structure of the hunting group and the type of prey. Read the free Plain Language Summary for this article on the Journal blog.

In situ observations of zooplankton show changes in abundance and swimming speed in response to hypoxia and acidification

AbstractZooplankton exhibit diverse swimming behaviors to reposition themselves in the water column, feed, find mates, and avoid predation. Environmental stressors that modify behavior can have cascading effects on population distributions and predator–prey interactions. Understanding zooplankton population dynamics is challenging, largely because traditional methods for quantifying zooplankton distributions are costly, limited in scope, and require extended analysis by trained analysts. We developed a novel methodology that combined remotely deployed camera systems, machine learning‐based identification of zooplankton, and video‐based tracking technology to quantify copepod and amphipod in situ swimming behavior in Hood Canal, WA, USA, a seasonally hypoxic and acidified fjord. Behavioral analysis showed copepods of all sizes swam on average 24% slower in stressful (hypoxic and acidified) waters relative to non‐stressful waters. Copepods exhibited less frequent escape responses in stressful waters, with a 68% decrease in the amount of time spent “jumping” for copepods 1–2 mm in length. Interestingly, abundances of small copepods increased in stressful waters, with 56% more 1–2 mm long copepods in stressful vs. non‐stressful conditions. In contrast, amphipods' average “darting” speeds did not differ between environmental conditions, but the abundance of amphipods significantly decreased in stressful waters relative to non‐stressful waters, suggesting avoidance of stressful conditions. Changes in swimming behavior are informative metrics in understanding ecosystem impacts of environmental stress because swimming speed has individual, population, and community‐level implications. Our results suggest that, among copepods, in situ behaviors may be useful proxies in monitoring the impacts of climate change on coastal ecosystems.

A sequential multi-agent reinforcement learning framework for different action spaces

In many real-world decision-making tasks, multi-agent need to learn collaboration in a high-dimensional complex action space, rather than just a single discrete action space. Recently, value decomposition learning methods such as QMIX have emerged as a promising approaches for collaborative multi-agent tasks. However, most of the value decomposition algorithms can only be used in discrete action space, which would limit their practicability. To address the limitation, we propose a novel algorithm called Multi-Agent Sequential Q-Networks (MASQN), which can be applied to the multi-agent domains with continuous, multidiscrete or hybrid action spaces. The proposed algorithm is based on the structure of centralized training with decentralized execution (CTDE). The decentralized actors ensure adaptability to different action spaces by utilizing action space discretization and sequential models, and the centralized critic utilizes the value decomposition architecture to guide effective updates of the policy parameters for each agent. We also give the convergence of joint policy from the perspective of policy iteration, by combining it with the CTDE structure and the constraint of the Individual Global Max (IGM) condition. Finally, we evaluate the MASQN algorithm on two benchmark environments: MAMuJoCo and Hybrid Predator–Prey. The empirical results show that MASQN out performs the state-of-the-art performance on three different action spaces.