Cats in farms: ranging behavior of the Fishing Cat (Prionailurus viverrinus) in a human-dominated landscape

Abstract

Home range studies provide valuable insights into animal ecology and behavior, informing conservation efforts and management strategies. Although the Fishing Cat (Prionailurus viverrinus) is a globally threatened habitat specialist species, only a few studies have been conducted on their home range and social organization, especially in response to human influence. In this study, we tracked 11 fishing cats with satellite GPS collars to investigate their home range size and habitat use in and around Koshi Tappu Wildlife Reserve, Nepal. The minimum convex polygon (MCP) and autocorrelation-informed kernel density estimation (AKDE) were used to estimate home range sizes of the fishing cats. Altogether 2,303 locations were obtained from 11 collared cats. The average home ranges of fishing cats (n = 8) with 95% MCP and 95% AKDE were 29.12 ± SD 16.89 km² and 39.88 ± 26.16 km², respectively. Home range (95% AKDE) of adult females (21.72 ± SD 16.39 km², n = 4) was significantly smaller compared to males (58.03 ± SD 21.19 km², n = 4). Sex-specific social organization with a single male overlapping with multiple nonoverlapping adult females was consistent with home range behaviors of other carnivores. The highest number of locations of collared fishing cats were in tall grasslands, whereas they highly preferred wetlands. A large part (over one-third) of fishing cat home ranges covers human-dominated areas such as fishponds, agriculture, and settlements encompassing various threats to fishing cats including persecution, road kills, and dog attacks. To ensure long-term survival of these cats amid habitat alteration and human–wildlife conflict, we recommend Fishing Cat conservation activities focusing on raising awareness, especially in human-dominated landscapes.

Fishing Cat, habitat use, home range, human-dominated landscape, wetland

The spatial analysis of home ranges and habitat use are key ecological parameters used to understand wildlife use of the landscape. In increasingly fragmented natural habitats, this information is necessary to identify areas that are important for survival of species and to design effective conservation plans, especially for cryptic carnivores which are nocturnal, shy, and camouflaged (Gehring and Swihart 2003; Nisi et al. 2023). The space requirements of a species also depend on factors including body size (Gittleman and Harvey 1982), availability of prey species, population density, sex, and human disturbances (Duncan et al. 2015). Home range studies can also help to identify potential conflicts between carnivores and humans such as predation on livestock or nuisance behavior, and information on home ranges can inform management strategies to mitigate these conflicts (Gehrt et al. 2009; Cutter 2015). Through advanced technologies such as satellite-enabled GPS collars, high-resolution data on space use patterns of species can be accurately obtained. However, conservation resources are limited, and research efforts are generally focused on large, charismatic rather than small, less charismatic species such that understanding of the spatial ecology of these species is limited (Brodie 2009; Macdonald et al. 2015; Tensen 2018).

Fishing cats (Prionailurus viverrinus) are globally “Vulnerable” with a patchy distribution associated with wetlands in South and South-East Asia (Mukherjee et al. 2016). They are known to have variable home range sizes (4 to 22 km²) depending on habitat and food availability (Sunquist and Sunquist 2002; Cutter 2015; Ratnayaka et al. 2022). Most studies on fishing cats focus on population status, and show that their population is patchily distributed and is decreasing throughout the range (Mukherjee et al. 2016). Only a few studies have used GPS collars to study fishing cats (Ratnayaka 2021). Space use and ranging behavior of fishing cats is not well-understood, yet is critical information for conservation management of this threatened species.

A large population of fishing cats is present in a human-dominated landscape outside protected areas (PAs) in countries including India, Bangladesh, and Nepal (Chowdhury et al. 2015; Kolipaka et al. 2019; Mishra et al. 2022a). There are a few published studies in Nepal that estimate the Fishing Cat population size and assess their distribution and threats (Mishra et al. 2018, 2021, 2022a). Scientific information on Fishing Cat movement, activity pattern, and home range in relation to habitat type is scarce for Nepal, and is mainly based on field-based very high frequency (VHF) radio tracking (Sunquist and Sunquist 2002). More advanced GPS-enabled satellite collars that allow regular, long-term tracking throughout the device’s lifetime have never before been used to monitor fishing cats in Nepal although the satellite tracking system has been used to obtain ecological information and improve the understanding of the species.

Satellite technology has revolutionized the collection of animal location data, especially in large, remote, or inaccessible areas (Habib et al. 2014). By attaching satellite tags to animals, we can gather precise geographic location data with time stamps that can be relayed remotely. This technology has significantly enhanced the understanding of animal ecology including movement, behavior, habitat selection, migrations, home ranges, human–wildlife conflict, and the impacts of climate change (Hebblewhite and Haydon 2010). Such data are crucial for effective conservation, informing PA plans, and assessing mitigation measures. Specifically, satellite collaring has been instrumental in estimating the home range of species, identifying core habitats, and supporting conservation action plans. However, it is relatively expensive to purchase and operate compared to VHF or ultra high frequency systems (data transmission costs). The technology faces limitations such as variable satellite coverage, fixed accuracy, and battery life, which can affect data quality and transmission (Kays et al. 2015). The data transmission from satellite collars needs high energy, which means short battery life compared to non-satellite tracking. Battery size cannot be increased beyond a certain limit (collar <3% of body weight) to ensure welfare of the animal. Thus, there is a trade-off between an increase in collar life with collection of fewer data points each day or recording high-resolution data, shortening the collar life (Kays et al. 2015). Additionally, capturing and tagging animals can cause distress and potential harm, necessitating the involvement of experienced personnel and veterinarians to ensure ethical and safe practices (Putman 2004).

The objectives of this study are: 1) to understand how fishing cats use their natural habitat inside the Koshi Tappu Wildlife Reserve and the surrounding human-dominated landscape in South-eastern Nepal; and 2) to understand their home range size and overlap. Most felid species exhibit a social organization wherein dominant males defend a territory that encompasses home ranges of several females (Sunquist 1981; Machado et al. 2017). We tested 2 hypotheses: 1) male fishing cats have a larger home ranges compared to females; and 2) fishing cats prefer both natural and man-made wetland habitats.

Materials and methods.

Study area.

This study was carried out in the Koshi Tappu Wildlife Reserve and its periphery in South-eastern Nepal (Fig. 1). The Reserve was established in 1976 with an area of 175 km² along the floodplain of Koshi River, a tributary of the Ganges River. An additional 173 km² area adjoining the Reserve was declared as a Buffer Zone in 2004. Koshi Tappu is a Ramsar site (wetland of international importance) and harbors many threatened species including wild Water Buffalo (Bubalus arnee), Gangetic River Dolphin (Platanista gangetica), Smooth-coated Otter (Lutrogale perspicillata), Swamp Francolin (Ortygornis gularis), and Bengal Florican (Houbaropsis bengalensis). While tigers (Panthera tigris tigris) have been locally extinct in the Reserve for a few decades, the Common Leopard (Panthera pardus) was recently reported in 2022 by the first author (RM) while camera trapping to monitor collared fishing cats. The Fishing Cat occurs along with other medium-sized carnivores including the Jungle Cat (Felis chaus) and Golden Jackal (Cuon alpinus). There is a high diversity of herbivores including Nilgai (Boselaphus tragocamelus), Chital (Axis axis), Hog Deer (Axis procinus), Barking Deer (Muntiacus vaginalis), Indian Hare (Lepus nigricollis), and rodents. Koshi Tappu has high fish diversity, with 141 species including 7 that are exotic (KTWR 2018). Commercial fish farming of 5 species including Bighead Carp (Hypophthalmichthys nobilis), Grass Carp (Ctenopharyngodon idella), Common Carp (Cyprinus carpio), Silver Carp (Hypophthalmichthys molitrix), and Nile Tilapia (Oreochromis niloticus) is practiced in the Buffer Zone.

Fig. 1.

Koshi Tappu Wildlife Reserve and its surrounding area in Nepal with Fishing Cat capture locations.

The main channel of the Koshi River causes a significant barrier for movement of animals, dividing the Reserve into Eastern and Western zones. The Koshi River is one of the highest sediment-transporting rivers in the world (Kafle et al. 2017), has extremely dynamic channels, and channel instability has been aggravated in recent years leading to excessive siltation that has resulted in the rise of the riverbed. Consequently, wetlands are being formed with seepage water along the outer side of the embankment and such areas have gradually been converted to commercial fishponds (Mishra et al. 2022b). There are fishponds in the villages as well.

The ponds at the edge of the Reserve attract fishing cats because of abundant prey including fish, amphibians, crustaceans, and mollusks. Presence of such predators around fish farms often leads to negative interaction with humans (Mishra et al. 2021; Chakraborty et al. 2022). Local communities also depend on the Reserve for irrigation, fishing, grazing, and non-timber forest products. In the past 4 decades, the wetland ecosystem has declined >30% which significantly affects numerous globally threatened species (Chettri et al. 2013; Chaudhary et al. 2016). In the fertile land of the eastern Buffer Zone of the Reserve, farmers grow different seasonal crops including sugarcane, maize, wheat, paddy, mustard, peas, sunflower, and jute which provides good cover to elusive species like fishing cats and jungle cats (Mishra et al. 2020).

In August 2008, the Koshi River breached the embankment at 500 m west of the Reserve headquarters and entered into agriculture and settlements (Kafle et al. 2017) converting 700 ha of fertile land to sand and water. Some part of this area is still covered by sediments or remains as marsh providing good habitat for the perennial herb Typha latifolia. Some part of the barren flooded land was converted into fish farms with Typha thickets around the edges. In some parts, where there remains year-round water forming the marshland with Typha, the Typha does not dry out, providing cover to fishing cats in the dry season as well, while thickets around the fish pond edges often dry up in winter and sprout again in summer. This combination of Typha and fish farms makes the habitat suitable for fishing cats and other species like otters. Most of the Typha is collected by the locals in winter to be used as fuel (mixed with dung) for cooking and knotting mats.

In the western part, natural riverine habitats occur with a mosaic of grassland patches, marshes, riverine forest, natural and man-made ponds (created by the Reserve authorities), smaller tributaries of rivers, and riverbanks. These habitats are rich with abundant prey, wetland birds, and fishes.

Fishing Cat capture and collaring.

We obtained research permission to collar the fishing cats from the Department of National Parks and Wildlife Conservation (DNPWC), Nepal. The research proposal was reviewed and approved by the technical committee of the Department. The capture and collaring of fishing cats were performed by trained veterinarians who adhered to the “Implementation Guideline for Satellite Telemetry on Fishing Cat (Prionailurus viverrinus) in Nepal, 2021” which was also approved by the DNPWC.

Between August 2021 and January 2023, a total of 11 fishing cats (6 females and 5 males) were fitted with iridium satellite collars manufactured by Africa Wildlife Tracking model AWT-IR SAT (Table 1). The collars have a lifespan of approximately 1 year. Metal box traps of size 42 × 15.5 × 16 in with fish baits were set in the evening at locations where fishing cats were likely to visit. Each trap was checked early in the morning and set open during the day hours. When a Fishing Cat was trapped in the morning, a trained veterinarian first administered a drug combination of ketamine (3 mg/kg) and medetomidine (0.07 mg/kg), and then performed a clinical checkup on the anesthetized cat. We recorded physical measurements including weight, body length, tail length, teeth, and paw dimensions of each captured cat.

Table 1.

Home range sizes of collared fishing cats using the MCP and AKDE methods. ID of the cat represents F—female, M—male, and SA—subadult.

Location	ID	Cat name	Weight (kg)	No. of active collar days	No. of locations used for home range estimation	MCP home range (km2)		AKDE home range (95% CI) (km2)
						MCP 95%	MCP 100%	50% AKDE	95% AKDE
Buffer Zone	F1	Gulabi	7.5	176	811	4.75	5.35	1.06 (0.95 to 1.17)	5.05 (4.53 to 5.60)
	F2 (SA)a	Jhalli	6.3	54	128	13.28	16.21	NA	NA
	M1	Kusaha	12	158	126	31.56	33.29	9.35 (7.78 to 11.06)	35.99 (29.96 to 42.57)
	F4 (SA)a	Kanchhi	6	96	154	1.16	1.67	NA	NA
	M2	Madhuban	10.5	192	104	38.11	46.91	17.42 (13.30 to 22.10)	63.35 (48.36 to 70.33)
	F3	Madhu	10	180	200	34.70	37.23	10.51 (8.31 to 12.97)	44 (34.77 to 54.29)
Reserve	M3	Pokhari	8	200	241	34.77	38.09	8.02 (5.65 to 10.79)	47.76 (33.67 to 64.28)
	F5	Pokhari	8	231	98	18.00	18.78	5.14 (4.17 to 6.22)	21.82 (17.69 to 26.37)
	M4	Triyuga	12	195	176	58.41	86.19	22.51 (19.30 to 25.97)	85.02 (72.89 to 98.07)
	F6	Triyuga	8	146	330	12.62	18.17	3.97 (3.49 to 4.47)	16.02 (14.10 to 18.06)
	M5a	Koshia	15	40	71	17.98	18.25	NA	NA

^aThese cats were only included in the MCP home range analysis and were excluded from AKDE home ranges as well as calculation of average home range sizes.

Once the cat was found healthy and suitable for collaring, it was fitted with a satellite collar and released after administering an antidote (atipamezole 0.21 to 0.35 mg/kg) within 30 to 45 min of sedation. The satellite collars were small (~200 g) and <3% of the weight of the animal, and as expected, we observed no adverse effects on the animals (Kenward 2000). To enhance the battery life we pre-programmed collars to obtain fixes (Kays et al. 2015).

The fishing cats were collared at 2 sites on the eastern and western banks of the Koshi River (Fig. 1) to ensure the coverage of a variety of potential Fishing Cat habitats. Six fishing cats (2 males, 2 adult females, and 2 subadult females <2 years) were collared in the Buffer Zone in the eastern part. Similarly, 5 fishing cats (3 males and 2 females) were collared within the Reserve in the western part. One female (F1) was caught twice to replace the collar. We classified cats as adults and subadults based on their body size (adults are >8 kg), teeth structure (subadults have sharp and white), and reproductive status (adult lactating females; Schroeder and Robb 2005).

The satellite collars were programmed to record and send location data every 3 h for the first 2 weeks. To save battery and maximize lifespan, they were then set to 2 to 6 locations per day. All data from the collars were downloaded through the Africa Wildlife Tracking web server and arranged in Excel spreadsheets.

Home range estimate and habitat use.

Two methods, the minimum convex polygon (MCP; Hayne 1949) and area-corrected autocorrelated kernel density estimation (AKDE; Fleming and Calabrese 2017), were used to estimate the home range sizes of fishing cats. We computed MCP home ranges including all locations obtained (100%) and 95% of the locations with fixed mean, respectively, using Home Range Tools in ARC GIS 10.4 (Rodgers et al. 2007). This method allows for comparison with the previous studies reporting only MCP home ranges. Similarly, the home range of each Fishing Cat was calculated using area-corrected AKDE (95% and 50% home range) using the “ctmm” package in R (Calabrese et al. 2016; Fleming and Calabrese 2017). This approach fits continuous time-movement models to the telemetry data and is efficient in dealing with the complexities of movement data which include autocorrelated data, small sample sizes, and missing or irregularly sampled data (Silva et al. 2022). The most likely model is selected by Akaike Information Criterion (AIC) and then used to estimate the home range area of an individual. Home range estimation using AKDE can overestimate range size for data of shorter time span (<3 months; Silva et al. 2022). To avoid such overestimates, we used only 8 out of 11 collared cats (discarded data of 3 cats having <3 months) having sufficient observations (98 to 811 locations during a 5- to 8-month period) for AKDE home range estimation.

We used a Poisson regression by constructing a generalized linear model (GLM; Zuur et al. 2009) to test the effects of sex and spatial location of the fishing cats on home range sizes. In the GLM, home range size (95% AKDE) was used as dependent variable. Two independent variables (sex and location—Reserve or Buffer Zone) and their interaction were used to examine their effect on home range size. Analysis was carried out in R using package “ctmm” (R Core Team 2016).

We also overlaid home ranges of different cats in ArcGIS 10.4 and calculated the areas of overlap with each other (Supplementary Data SD1 and SD2). Welch’s t-test was used in R to compare the home range size of males and females as well as natural habitat (Reserve) and human-dominated areas (Buffer Zone).

We investigated habitat selection by fishing cats using Ivlev’s electivity index (Ivlev 1962), Jacob’s index (Jacobs 1974), and habitat selection ratios (Kauhala and Auttila 2010). Habitat type was obtained from the global land cover data at 10-m resolution produced from Santinel-2 satellite images (Karra et al. 2021). Fishing Cat locations were overlaid on the land cover map, and the proportion of locations within each habitat was calculated for daytime (06:00 to 18:00), nighttime (18:00 to 06:00), and all day combined. The proportion (as percentage) of habitat available within the study was calculated overlaying the MCP of all collared cats on the land cover map. The Ivlev’s index E_i for habitat i is calculated with the formula: E_i = (u_i − a_i)/(u_i + a_i), where u_i is the proportion of fishing cat locations in habitat i (habitat utilized), and a_i is the proportion of habitat i available in the study area. Similarly, the Jacobs index (D) was calculated using the formula: D = (r − p)/(r + p − 2rp), where r is the proportion of habitat used and p is the proportion of habitat available. The value of E_i and D ranges between −1 (strong avoidance) and +1 (strong preference)—values close to 0 indicate the use of habitat proportionately to its availability. Moreover, the selection ratio (% habitat used/% habitat available) for each habitat type was also calculated (Kauhala and Auttila 2010).

Results

We obtained a total of 2,664 locations from 11 collared fishing cats. We excluded 361 locations from the analysis that were repeated overlapping records within an hour. So, altogether 2,303 locations from all 11 collared cats with locations ranging between 71 and 811 per Fishing Cat, operational for 40 to 231 days, were analyzed for MCP home range (Table 1). The AKDE home range was estimated for 8 cats with more than 3 months of data.

An average home range of fishing cats (n = 8) with 95% MCP was 29.12 ± SD 16.89 km² and 95% AKDE was 39.88 ± 26.16 km² (Table 2). Home range (95% AKDE) of adult females (21.72 ± SD 16.39 km², n = 4) was significantly smaller (t = −2.71, df = 5.64, P = 0.037) compared to the males (58.03 ± SD 21.19 km², n = 4). There was no overlap of home ranges between adult females both in the Wildlife Reserve and Buffer Zone. Core area of the home range (50% AKDE) of all cats was 9.75 ± SD 7.13 km², whereas males have 11.46 ± SD 6.86 km² and females have 5.17 ± SD 3.95 km². There was high overlap of the home ranges of males with that of females, and between males—but there was no overlap between adult females (Table 2; Fig. 2; Supplementary Data SD1 and SD2). An average home range (95% AKDE) of fishing cats inside the Reserve (42.66 ± 31.43 km², n = 4) did not differ significantly (t = 0.27, df = 5.63, P = 0.78) compared to the Buffer Zone (37.10 ± SD 24.25 km², n = 4). The Poisson regression analysis showed that sex was an important factor explaining Fishing Cat home range size. Location only (Reserve or Buffer Zone) had no significant effect on home range size but the effect was significant with interaction of location (Reserve) and sex (male). The male fishing cats had significantly larger home ranges compared to females (Table 3).

Table 2.

Summary of the home range sizes of fishing cats, grouped by sex and location and overlap between them.

Sample population	Mean MCP home range (km2)		Mean AKDE home range (km2)
	MCP 95% ± SD	MCP 100%	50% AKDE ± SD	95% AKDE ± SD
All fishing cats (n = 8)	29.12 ± 16.89	35.5 ± 24.50	9.75 ± 7.13	39.88 ± 26.16
Females (n = 4)	17.52 ± 12.68	19.89 ± 13.12	5.17 ± 3.95	21.72 ± 16.39
Males (n = 4)	40.71 ± 12.10	51.12 ± 24.05	11.46 ± 6.86	58.03 ± 21.19
Cats in Buffer Zone (n = 4, eastern part)	27.78 ± 15.26	30.70 ± 17.83	9.58 ± 6.71	37.10 ± 24.25
Cats inside Wildlife Reserve (n = 4, western part)	30.95 ± 20.59	40.31 ± 31.95	9.91 ± 8.57	42.66 ± 31.43

Overlap between	Male/Female	95% MCP (km2)		95% AKDE (km2)
Western part
Pokhari F5 & Triyuga F6	Female–Female	0		0
Pokhari F5 & Pokhari M3	Female–Male	2.54		9.64
Pokhari F5 & Triyuga M4	Female–Male	15.81		15.72
Triyuga F6 & Pokhari M3	Female–Male	5.7		8.81
Triyuga F6 &Triyuga M4	Female–Male	4.4		9.63
Triyuga M4 & Pokhari M3	Male–Male	0.94		6.68
Eastern part
Gulabi F1 & Madhu F2	Female–Female	0		0.01
Gulabi F1 & Kusaha M1	Female–Male	2.82		3.04
Gulabi F1 & Madhuban M2	Female–Male	4.75		5.06
Madhu F2 & Kusaha M1	Female–Male	0		0.22
Madhu F2 & Madhuban M2	Female–Male	11.59		19.21
Kusaha M1 & Madhuban M2	Male–Male	14.14		25.71

Table 3.

Parameter values of individual variables of GLM fitted to Fishing Cat home range size in Koshi Tappu Wildlife Reserve (HR ~ Sex + HR.coverage + Sex × HR.coverage, family = Poisson).

Parameters	Estimate	Std. error	z-value	Pr(>|z|)
(Intercept)	3.1987	0.1429	22.391	<0.001***
SexM	0.7033	0.1747	4.026	<0.001***
HR.coverage—Reserve	0.2542	0.2162	−1.176	0.198
HR.coverageReserve:SexM	0.5495	0.2537	2.166	0.0303*

^*P <0.05, ***P <0.001.

Fig. 2.

Locations of individual fishing cats (name and ID number) in Koshi Tappu Wildlife Reserve and its Buffer Zone. Each contour represents AKDE 95% home range of Fishing Cat individuals.

Habitat use of the Fishing Cat occurred in the order: tall grassland/shrub (31.51%), wetland (27.36%), agriculture (23.95%), forests (8.68%), settlements (5.02%), bare ground (2.45%), and short grasslands (1.04%; Fig. 3). However, the Ivlev’s electivity index, Jacob’s index, and habitat selection ratio show a strong preference of fishing cats for wetlands over all other habitat types both during day and nighttime (Table 4). Fishing cats used wetlands and settlements more during nighttime compared to daytime, whereas tall grasslands were used more during daytime.

Table 4.

Habitat selection of Fishing Cat in Koshi Tappu Wildlife Reserve, Nepal.

Habitat type	Use (%)	Available (%)	Ivlev’s electivity index	Jacob’s index	Selection ratio
Overall (day and night) locations combined
Bare ground	2.45	6.36	−0.44	−0.46	0.39
Short grassland	1.04	4.28	−0.61	−0.62	0.24
Wetland	27.36	6.42	0.62	0.69	4.26
Agriculture	23.95	29.39	−0.10	−0.14	0.81
Tall grasslands & shrubs	31.51	37.09	−0.08	−0.12	0.85
Settlements	5.02	5.57	−0.05	−0.05	0.90
Forests	8.68	10.88	−0.11	−0.12	0.80
Nighttime locations
Bare ground	2.78	6.36	−0.39	−0.46	0.44
Short grassland	1.23	4.28	−0.55	−0.62	0.29
Wetland	29.05	6.42	0.64	0.69	4.53
Agriculture	23.22	29.39	−0.12	−0.14	0.79
Tall grasslands & shrubs	29.70	37.09	−0.11	−0.12	0.80
Settlements	6.10	5.57	0.05	−0.05	1.09
Forests	7.92	10.88	−0.16	−0.12	0.73
Daytime locations
Bare ground	1.30	6.36	−0.66	−0.46	0.20
Short grassland	0.37	4.28	−0.84	−0.62	0.09
Wetland	21.48	6.42	0.54	0.69	3.35
Agriculture	26.48	29.39	−0.05	−0.14	0.90
Tall grasslands & shrubs	37.78	37.09	0.01	−0.12	1.02
Settlements	1.30	5.57	−0.62	−0.05	0.23
Forests	11.30	10.88	0.02	−0.12	1.04

Fig. 3.

Percentage of locations of each Fishing Cat in different habitat types in Koshi Tappu Wildlife Reserve, Buffer Zone, and adjoining areas.

Discussion

Our study conducted the first-ever GPS satellite collar field research of fishing cats in Nepal. This study was conducted simultaneously within a PA (Wildlife Reserve) and outside (human-dominated landscape) with a sample size of 11 cats to understand their home range size and habitat use. We documented a larger average home range size of fishing cats compared to values previously reported (Table 5). Male fishing cats covered significantly larger home ranges compared to females. The average male and female fishing cat home range sizes in natural habitats of core PAs (rivers, grasslands, and forests) were also larger compared to those in the human-dominated landscape (agriculture, fish farms, and settlements). In terms of habitat preference, we documented a preference for wetland habitats.

Table 5.

Summary of Fishing Cat home range estimates across their range.

Location	Type of habitat	Study period	Sample size (sex and number of adult individuals)	Equipment and estimation method used	Average home range area in km2 (M = male, F = female)	Source
Chitwan National Park, Nepal	PA	1984	Male—1 Females—3	VHF collars, MCP 95%	M: 16 to 22	Sunquist and Sunquist 2002
Colombo, Sri Lanka	Urban area	2013 to 2015 2018 to 2019	Males—2 Females—2	GPS collars, time local convex hull 95%	M: 12.49 F: 1.17	Ratnayaka 2021
Khao Sam Roy Yot National Park, Thailand	PA and agriculture area	2009 to 2010	Males—2 Females—4	VHF collars, fixed kernel density 95%	M: 8.75 F: 3.85	Cutter 2015
Krishnapur Wildlife Sanctuary	PA	2011 to 2012	Male—1	Camera trap, MCP 100%	M: 18.46	Nair 2012
Koshi Tappu Wildlife Reserve	PA and agriculture area	2021 to 2022	Male—4 Female—4	Satellite GPS collar, autocorrelation-corrected kernel density 95%	M: 58.03 F: 21.72	This study

There are only a few studies on Fishing Cat spatial ecology with limited information on their home range sizes in different habitats (Table 5). Some previous studies show that comparable estimates of the home range size of fishing cats, for example, in Chitwan National Park and Kishanpur Wildlife Sanctuary, could be attributed to similar socioecological characteristics of habitats (Sunquist and Sunquist 2002; Nair 2012). However, the home range estimate of fishing cats in our study was much larger than the home range sizes reported in previous studies (Table 5).

There may be 2 reasons for the relatively larger home range size of fishing cats in our study compared to other studies. First, the automatic location recording by GPS collars ensures sampling of the full coverage of the areas used by fishing cats with an adequate number of locations, whereas VHF telemetry often fails to locate fishing cats, especially when the cats are not within the VHF receiver range, thereby often underestimating their home range size. Second, habitat characteristics of the study area might include lower prey availability compared to other studies. Only a study in Sri Lanka (Ratnayaka et al., 2021) used GPS collars with the ability to automatically record multiple GPS locations in a day, providing a more accurate account—all other studies used yagi antennas for collecting coordinates. The smaller home range size of fishing cats in Sri Lanka (Colombo) compared with those in our study may be explained by higher prey densities of rodents in the urban areas of Colombo city (Ratnayaka et al. 2022)—duration of monitoring was also short (mean 127 days, n = 5) compared to our study (mean 185 days, n = 8).

The average home range size of fishing cats living in the eastern part (Buffer Zone) was slightly smaller than the western part (Reserve) of the Koshi River. We documented relatively smaller home ranges of fishing cats living in agriculture and fish farms of the Buffer Zone area compared to those living inside the core areas of the Reserve. The abundance of prey (fish in the fish ponds and rodents in agriculture fields) is probably higher in the human-dominated areas compared to core areas within the Reserve. This difference may also result from depletion of fish stock due to a high fishing pressure in the natural wetlands and Koshi River inside the Reserve.

Our result of significantly larger home range sizes of male compared to female fishing cats is consistent with previous studies conducted in Nepal, Thailand, and Sri Lanka (Sunquist and Sunquist 2002; Cutter 2015; Ratnayaka et al. 2022). In carnivores, males generally try to maximize the coverage of females within their home ranges to increase the chances of breeding, whereas females usually select areas with abundant prey and safe shelters for raising the cubs (Wolff and Peterson 1998). Similar observations of larger home ranges of males overlapping with multiple females have been reported in other carnivores (Sunquist 1981). The home range of 1 breeding female (F1) was comparatively smaller than other adult females.

We documented a large overlap of male home range with females and some overlap with other males. Unlike males, there was no overlap between adult females. Although it is believed that males are more aggressive in defending their territory, Fishing Cat females seem to maintain their exclusive territories. However, we also documented some sharing of habitat by adult females (F1 and F3) with other subadult females (F2 and F4) who are probably their daughters. For example, the 2 females F1 and subadult F2 were recorded together (in a same frame) in camera traps close to the location where F1 was collared. F2 was covering twice the area of F1, likely in the process of separating from F1 and looking for her own territory.

Though our estimates of fishing cat home ranges are quite large, and fishing cats have strong swimming abilities to cover longer distances, none of the collared cats crossed the main channel of Koshi River. This finding suggests that they avoid crossing deep or high-current water bodies, or that there is sufficient prey that outweighs the risky effort of crossing the river. Our study was conducted during the dry season with low water volume in the river.

Fishing cats are known to be associated with wetland habitats (Mishra et al. 2018), which is also confirmed by our study. Only the wetlands had a positive value (identified as preferred habitat) in Ivlev’s electivity index. Although wetlands were the most preferred habitat, fishing cats were found more frequently (nearly one-third of the locations obtained from collars) in the tall grasslands and shrubs compared to other land cover types, which is expected because this habitat covers a large portion of the study area (~56%; Chaudhary et al. 2016). A similar observation of fishing cats using tall grasslands close to wetlands has been reported from Chitwan National Park (Mishra et al. 2018). Interestingly, we also documented fishing cats using settlements and agriculture areas (~30% of the locations). A similar observation of fishing cats using urban and highly altered areas has been reported from Sri Lanka (Ratnayaka et al. 2022). We documented higher use of agriculture and tall grassland, whereas lower use of wetlands by fishing cats during daytime compared to nighttime—they might be taking shelter in these habitats during the day and visiting wetlands in search of prey during night.

Based on the limited information obtained from a camera trap study in Koshi Tappu, Mishra et al. (2021) assumed that fishing cats visit the fishponds in the Buffer Zone during the night and remain inside the Reserve for cover during the day. However, except for 1 male (M1), we observed that fishing cat home ranges are either entirely or mostly situated in the Buffer Zone. The 4 females collared in the Buffer Zone were taking shelter in bushes around the fish ponds (typically Typha spp.), sugarcane field, and paddy as per availability. In the Buffer Zone, fishing cats primarily used the Typha field throughout both day and night during monsoon and post-monsoon seasons when the plants are dense and tall. After the Typha plants dry, local people harvest them for household use/handicrafts or burn them for cleaning the fields and for better sprouting in the next season. During this period, fishing cats switch their shelter to the sugarcane field or other crops for cover during the day, but they still visit the fish ponds at night. We conclude that such adaptability contributes to their success in human-dominated landscape.

However, there are multiple threats to fishing cats in human-dominated landscapes including retaliation by humans after conflict (taking fish from fish ponds), road kills, and attacks by guard dogs (Mishra et al. 2021). For instance, one of the subadult females (F3), who remained exclusively in the Buffer Zone, was found dead after 3 months of collaring due to injury in the neck—she was probably attacked by a guard dog. A previous case of a Fishing Cat being killed by dog has been reported from the Buffer Zone (Mishra et al. 2021). The body of adult male M5 was recovered in the southern Buffer Zone of the Reserve after 40 days of collaring---is ribs were broken and we suspect that he was beaten to death by locals. The adult female F1 was also stuck in the drainage pipe in a paddy field after 2 months of collaring—fortunately, she was successfully rescued and released and was recorded with a kitten (5 to 6 months old) in a camera trap 7 months after this incident indicating that she was pregnant at the time and conceived after her collaring. Thus, despite intensive use of the Buffer Zone habitats, Fishing Cat survival is challenging amid escalating threats.

In conclusion, based on our findings, Fishing Cat conservation activities should focus equally on natural habitats and human-dominated landscapes. To ensure long-term survival of these cats amid habitat alteration and human–wildlife conflict, we recommend Fishing Cat conservation activities focusing on raising awareness especially in human-dominated landscapes and conflict mitigation measures.

Supplementary data

Supplementary data are available at Journal of Mammalogy online.

Supplementary Data SD1. The home range overlapping areas among collared fishing cats using 100% MCP.

Supplementary Data SD2. GPS locations of individual fishing cats (name and ID) in Koshi Tappu Wildlife Reserve and its Buffer Zone. Polygons represent minimum convex polygon (MCP) 100% home range areas.

Acknowledgments

We thank the Nepal Department of National Parks and Wildlife Conservation and the Koshi Tappu Wildlife Reserve Office for granting research permission. We appreciate the support from the University of Antwerp, the National Trust for Nature Conservation, and WILD CARE, Nepal. Our sincere thanks to Anya Ratnayaka and Jim Sanderson for their continuous advice during the Fishing Cat collaring, and partners of Fishing Cat Conservation Alliance for their support. Collaring of the cats was supported by veterinarians Puroshottam Pandey, Purushottam Mudbary, and Kiran Rijal. We acknowledge the support of Koshi Tappu Wildlife Reserve staff including Chief Conservation Officer Ashok Kumar Ram and staff of the National Trust for Nature Conservation–Koshi Conservation Center including Birendra Gautam, Tika Ram Tharu, and Ohm Prakash Chaudhary. Bishnu Lama provided significant technical support for trapping and collaring. We are grateful to the support of Hadrien Haupt of African Wildlife Tracking for helping us in setting the commands to satellite collars when required. We acknowledge the support of Dave Johnson from Katie Adamson Conservation Fund. We thank Amy Zuckerwise and 2 anonymous reviewers for review and critical suggestions in the manuscript. We are thankful to local fish farmers Gulabi Mukhiya and Paresh Bista, who allowed us to capture fishing cats on their fish farms and install camera traps for monitoring. Local residents Jageshwor Yadhav and Surendra Pandit also supported us during the fieldwork.

Author contributions

RM, HL, BRL, NS, HRA and HHdI designed the study. RM and SA performed the fieldwork. BRL and RM analyzed the data. RM wrote the original draft. All reviewed and edited the manuscript.

Funding

Funding was provided by the Schlumberger Foundation—Faculty for the Future Fellowship program, the Leo Foundation, The Netherlands, and the Small Wild Cat Conservation Foundation, United States.

Conflict of interest

None declared.

Data availability

Data and code to replicate our results are available upon request.