Ichthyofauna of the Kubo, Tochikura, and Ichinono river systems (Kitakami River drainage, northern Japan), with a comparison of predicted and surveyed species richness

Abstract The potential fish species pool of the Kubo, Tochikura, and Ichinono river systems (tributaries of the Iwai River, Kitakami River drainage), Iwate Prefecture, northern Japan, was compared with the observed ichthyofauna by using historical records and new field surveys. Based on the literature survey, the potential species pool comprised 24 species/subspecies but only 20, including 7 non-native taxa, were recorded during the fieldwork. The absence during the survey of 11 species/subspecies from the potential species pool suggested either that sampling effort was insufficient, or that accurate determination of the potential species pool was hindered by lack of biogeographic data and ecological data related to the habitat use of the species. With respect to freshwater fish conservation in the area, Lethenteron reissneri, Carassius auratus buergeri, Pseudorasbora pumila, Tachysurus tokiensis, Oryzias latipes, and Cottus nozawae are regarded as priority species, and Cyprinus rubrofuscus, Pseudorasbora parva, and Micropterus salmoides as targets for removal.


Introduction
Biodiversity is rapidly declining at a global level due to a variety of anthropogenic pressures (Myers et al. 2000, Roberts et al. 2002, Brooks et al. 2006). Japan's biodiversity is no exception, primarily because of the country's high human population density and its status as a biodiversity hotspot with high endemism, including among freshwater fishes (Boufford et al. 2004, Brooks et al. 2006. Against this background, activities for the conservation and/or preservation of freshwater fishes have expanded in many regions of Japan. These include a number of nature restoration projects based on the Law for the Promotion of Nature Restoration, a Japanese domestic law that took effect in 2003 (Fujimoto et al. 2008, Matsuzaki et al. 2011.
Understanding a region's biota is critical for successful nature restoration projects, particularly with respect to biodiversity conservation. When detailed survey data are lacking, determining a region's potential species pool (Lessard et al. 2012) can provide an estimate of local biodiversity status by application of the ecological filter concept (Fattorini and Halle 2004, Hobbs and Norton 2004, Funk et al. 2008. Based on this concept, species diversity at the local scale is presumed to be constrained by the regional species pool, and therefore filtered by stochastic and deterministic processes including both nonanthropogenic and/or anthropogenic factors. Such factors exert pressures on regional species diversity over longer time frames than more recent influences such as the present biodiversity crisis. An understanding of the biogeography of a region can allow us to compile a list of potential species that provides baseline data for the evaluation and design of an action plan for regional biodiversity conservation and restoration. The nature restoration committee of the Kubo-gawa Ihatov area represents one of the projects based on the Law for the Promotion of Nature Restoration, but is unique in being overseen by private sector administration (Chisaka 2010). Part of this project has focused on the conservation and restoration of the freshwater fishes of the region, which includes most of the Kubo, Tochikura and Ichinono river systems.
In this study, we consulted historical literature to determine the potential species pool of this region (i.e., the species we predict to occur there), and carried out comprehensive field surveys to "ground truth" this list, as well as to identify fish species of conservation and restoration concern.

Materials and methods
The study region is located in the Kubo, Tochikura and Ichinono river systems (38°52′-38°5 5′ N, 140°56′-141°03′ E; Fig. 1), which are all tributaries of the Iwai River within the Kitakami River drainage that empties into the Pacific Ocean (Ichinoseki City, Iwate Prefecture, northern Japan). It includes the headwaters of these rivers, and headwater areas in smaller valleys (Fig. 1). The Kubo and Tochikura rivers each have a sediment control dam constructed in the lower reaches (Fig. 1). Rural cultivation and the typical ecological networks created by rice paddy ecosystems (traditional secondary nature, " satoyama" landscape: Washitani 2008) are widely established in the region (Ohtani et al. 2013;Fig. 2a).
The potential fish species pool of the study region was determined by a review of existing literature (Kawanabe et al. 2001, Ozawa et al. 2003, Matsuzawa and Senou 2008, Kumagai and Katsushima 2012, Nakabo 2013 as follows: the freshwater fishes with a natural distributional range located entirely or partly within Iwate Prefecture were catalogued, and the subsequent list was refined by considering ecology of the individual species and geography of the study region (i.e., whether they are known to inhabit headwaters and/or upper reaches of Japanese rivers or not). In other words, the fish species included in the potential species pool were determined by literature reports recording the species from the headwaters and/or upper reaches of rivers in Japan. For example, Kawanabe et al. (2001) reported that the Japanese eel, Anguilla japonica, often inhabits the upper reaches of Japanese rivers including Iwate Prefecture, and therefore we included A. japonica in our potential species pool of the region.
The fish fauna was assessed during field surveys using hand nets, minnow traps, cast nets, set nets, and hook-and-line angling. More than 200 irrigation ponds and three sampling sites (upper, middle and lower reaches) in each of the three rivers were sampled from April to October in 2008October in , 2009October in and 2010 (Fig. 1). The substratum of the irrigation ponds consisted mainly of silt and organic litter with aquatic plants (Fig. 2b, c). The Map of the Kubo, Tochikura, and Ichinono river systems. Black circles and red trapezoids indicate the main survey points where fish sampling was conducted and sediment control dams, respectively.
Voucher specimens and photographs are deposited in the Kanagawa Prefectural Museum of Natural History, Odawara, Japan (KPM) and the National Museum of Nature and Science, Tsukuba, Japan (NSMT). Unless otherwise stated, the systematic arrangement of families and scientific names follows Eschmeyer (2014), while that of standard Japanese names follows Nakabo (2013).
To evaluate the accuracy of the potential species pool (the predicted ichthyofauna), we made rarefaction curves based on presence/absence data from our field surveys. This allows us to estimate the number of additional species that our surveys may have missed. These analyses were conducted using EstimateS with the bias-corrected formula (Chao et al. 2004).

Distribution: Japan
Notes: This taxon is identical to Carassius buergeri subsp. 2 of Hosoya (2013). Hosoya (2002) treated this species as a subspecies of C. auratus based on genetics, ecological features and comparison with the types in RMNH (but this work is still under preparation). This taxon was mostly collected from lentic environments, but occasionally from lotic environments.

Distribution: Japan
Notes: We do not follow the Catalog of Fishes, but follow Kottelat (2006) in the nomenclature of this species. This taxon is identical with Phoxinus lagowskii steindachneri of Hosoya (2013), and was recorded from lotic waters and the inlets to some lentic waters.

Distribution: Japan.
Notes: This taxon is identical to Pseudorasbora pumila pumila of Hosoya (2013), in light of a potential new subspecies of Pseudorasbora pumila being recognized from Nagano, Shizuoka, Aichi, and Gifu Prefectures (Hosoya 2002, Hosoya 2013

Distribution: Japan.
Notes: This species usually inhabits lotic environments of river (Kawanabe et al. 2001, Hosoya 2013. However, Sugiyama (1997) mentioned that this species often inhabits isolated ponds in the mountains in Tohoaku region. Our records of this species also include the isolated irrigation ponds as well as the rivers. Notes: Japanese researchers have considered that the Rhinogobius brunneus complex currently includes many valid, synonymous species,and potentially several undescribed species (Akihito et al. 2002, Akihito et al. 2013. This taxon of this region matches the "Rhinogobius sp. OR of Akihito et al. (2002)" species complex (see Akihito et al. 2002, Akihito et al. 2013. It was recorded only from lentic waters including irrigation ponds, channels, and slow moving parts of rivers in the present study.

Analysis
Following the literature review, 35 freshwater fish taxa, including subspecies, belonging to 12 families in eight orders were considered to occur naturally in Iwate Prefecture (Table 1). Of these, 24 species/subspecies were determined to constitute the potential species pool of the survey area following our literature review (Table 1). List of fishes occurring naturally in Iwate Prefecture, northern Japan. From this list the potential species pool was determined by the literature review (ref. Kawanabe et al. 2001, Ozawa et al. 2003, Matsuzawa and Senou 2008, Kumagai and Katsushima 2012, Nakabo 2013. CR, critically endangered; EN, endangered; VU, vulnerable; NT, near threatened; DD, data deficient; LC, least concern; LP, threatened local population; category of red lists (Ministry of the Environment, Japan 2013, Iwate Prefecture 2014). YES, NO; Listed as non-native species (Matsuzawa andSenou 2008, Nakabo 2013); Record from each study region based on the field surveys.  The rarefaction curves of survey data from irrigation ponds and rivers are shown in Fig. 8; the expected saturation point of species richness was reached.

Discussion
Only 13 of the 24 species/subspecies from our estimated potential species pool were collected during field surveys, including nine national/prefectural red list species (Table 1). In addition, seven non-native species, including one national/prefectural red list species and three non-native invasive species, were recorded (Table 1). a b c d Figure 7.
Fishes collected from the Kubo, Tochikura, and Ichinono river systems (continued). Ten species/subspecies of the potential species pool for the survey area were not recorded during field surveys (  , 1907, andCottus pollux Günter, 1873) were considered to have a low probability of occurrence in the study region due to a limited distribution in a b Figure 8.
Rarefaction curves analyzed using presence/absence data of native fish species. The solid line indicates the mean value, while the dashed lines indicate the 95% CI upper and lower limits.
a: at the irrigation ponds b: at the rivers Tohoku region (they depend on sea currents for their migration and/or were near the boundaries of their distributional ranges) (Kawanabe et al. 2001, Ozawa et al. 2003, Nakabo 2013. It is also possiblethat our predicted species pool for the survey region was inaccurately estimateddue to a lack of data regarding biological interactions, heterogeneity of environments, and quality/quantity of the various habitats. In other words, some species included in the potential species pool might well have been excluded if we had more data on their ecology, particularly with respect to their biological interactions with other fishes. There is also the chance of local extirpation. The possibility that insufficient sampling efforts influenced the results of the present study is supported by the rarefaction curve as described below. The curve analyzed by using the presence/absence data of the irrigation ponds suggests that no more native species will be recorded from the ponds of the region due to the curve reaching saturation (Fig. 8a). However, the rarefaction curve of the rivers suggests that 1-3 additional nativepecies could be recorded based on the curve's saturation point and its 95% CI upper bound (Fig. 8b). It is also important to state that the species richness estimates assume that our survey techniques have an equal probability of capturing every species, but this assumption is likely to be violated in practice due to the different characteristics of each species (i.e. small benthic species or large active species maybe harder to catch).
The Kubo and Tochikura Rivers are both dammed, and Zacco platypus, Pseudogobio esocinus esocinus, and Oncorhynchus keta were not recorded upstream of the dams. This suggests that the dams have negatively affected migratory behavior in these species (although Z. platypus is a naturalized non-native species).
The meta-population of Oryzias latipes in the study region was possibly introduced by local residents. The altitude at which it was recorded (105-160 m) is perhaps too high to consider this a natural occurrence. Populations of O. latipes in Kanagawa Prefecture naturally occur at a height of~30 m above sea level (Senou unpublished), for example. Additionally, the specimens collected did not exhibit breeding condition, except at one sampling site (Miyazaki 2010). According to the local residents, these non-breeding populations probably all originated from the same source near the river system. This species is highly threatened and locally extirpated in many parts of Iwate Prefecture, including elsewhere in the river system studied here, and more generally in Japan (Ministry of the Environment, Japan 2013, Iwate Prefecture 2014). Therefore, we regard these populations of O. latipes as a target for ex-situ conservation because of the high vulnerability of the original populations and the less damaging ecological effects in the study region.
Although we tentatively identified all specimens of the genus Carassius, except for C. cuvieri, as C. auratus buergeri, the specimens display some morphological variability. One specimen (KPM-NI 23745: Fig. 3e) has 16 soft, branched dorsal-fin rays (vs. 12-14 in C. a. b.), and the other specimens have an angular jugular (lower jaw). Within Japanese crucian carps, the angular jugular is one of the specific characters of Carassius auratus grandoculis Temminck & Schlegel, 1846 (Hosoya 2013), whereas Takahashi (2003) reported that C. a. b. populations from irrigation ponds of Ehime Prefecture tend to have an angular jugular. Further information on the morphology, molecular biology, and biogeography of Japanese crucian carps is urgently needed, because they include national and regional red list species and/or populations (Ministry of the Environment, Japan 2013; Iwate Prefecture 2014). In fact, Dr. Nakamura, who published the current taxonomy of Japanese crucian carps (Nakamura 1969), indicated that morphological variations and distributional ranges of C. a. b. in northern Japan are in need of clarification (Nakamura 1969, Nakamura 1976).
Among the red list species recorded in the present study, Lethenteron reissneri, Pseudorasbora pumila, Tachysurus tokiensis, Oryzias latipes, and Cottus nozawae (threatened species) were rare and are considered vulnerable. Further research of these species is required, particularly with regard to their population dynamics and conservation in the Kubo-gawa Ihatov area. Similarly, the population dynamics and extended distributions of the non-native and invasive species should be studied in greater detail with a special focus on potential methods of removal. In particular: Cyprinus rubrofuscus, which has been reported as having highly negative effects on native macrophytes and odonates in the region (Miyazaki et al. 2010, Sekizaki et al. 2012, Yoshioka et al. 2014; Pseudorasbora parva, which has replaced P. pumila following hybridization in many localities (Konishi andTakata 2004, Konishi 2010); and Micropterus salmoides, which has apparently been illegally introduced by sport fishermen and/or residents (Miyazaki 2010, Tsunoda et al. 2011).
We refer the above suggestions to the focal nature restoration committee (see also Chisaka 2010). In other words, the localities of vulnerable species (i.e., Lethenteron reissneri, Carassius auratus buergeri, Pseudorasbora pumila, Tachysurus tokiensis, Oryzias latipes, and Cottus nozawae) will be preferentially conserved and that of nonnative species that negatively affect sustainability of native species (i.e., Cyprinus rubrofuscus, Pseudorasbora parva and Micropterus salmoides) will be preferentially removed. In addition, we will further investigate suitable habitats in the target area to record the remaining uncaptured species/subspecies from the potential species pool. Therefore, the comparison of the potential species pool with the national and regional Red Lists provides useful suggestions for priority species that should be conserved or removed, and for determining further sampling sites with suitable habitat for the absent species (those listed in the potential species pool that were not yet collected in the field surveys).
The determination of the freshwater fish species pool of a given region is an essential first step in restoration ecology (Fattorini and Halle 2004, Hobbs and Norton 2004, Funk et al. 2008. When followed up with field surveys, desk surveys can lead to an improved understanding of regional biotas and their primary characteristics,permit prediction of potential extinction events caused by natural or anthropogenic factors, and provide insights regarding the geological history affecting freshwater fish distributions. Furthermore, many restoration projects are somewhat restricted by budgets, human resources, and/or methodology (e.g., Joseph et al. 2009, Carwardine et al. 2012, however, desk surveys are required as an important first step to determine the potential species pool. However, this information can be limited, as shown here, and precise information from field surveys can then feed back into improved estimates for future desk surveys and restoration ecology projects.