BACKGROUND It is of great significance to study the breeding environment of the naked carp in Qinghai Lake for the conservation and restoration of natural habitat. The naked carp is the only economic species in Qinghai Lake, colloquially called “Huang Yu”, and plays a core role in the lake ecosystem. As an anadromous species, the naked cap migrates between Qinghai Lake and major rivers to spawn such as the Buha, Shaliu, Quanji, Heima, and Hargai rivers from April to August every year. Due to intensive fishing, arid climate, and the lack of spawning rivers, the amount of the naked carp’s resources has dropped sharply, from 28000 tons in the 1960s to 2263 tons in the 1990s. In view of its important ecological status, it is urgent to restore and protect the resources of naked carp in Qinghai Lake. Since 1982, Qinghai Province has forbidden fishing to restore and protect the fish resources. At the same time, it has established fish conservation bases and artificial breeding and releasing stations around Qinghai Lake. By 2022, the fish resources of the naked carp had increased to 108500 tons, 42 times as many as in the early stages of protection. The breeding environment is the key factor for the continuation of the fish population, and different fish require different parameters of the spawning ground, such as river, light, sediment and water temperature. It is also very important to obtain the reproductive environmental parameters and determine the optimal spawning ground for the protection and expansion of the fish resources. At present, only Zhou et al.³ have discussed the environmental requirements of natural reproduction of the naked carp by indoor artificial simulation. Therefore, it is of great significance to study the breeding environment of the naked carp for the conservation and restoration of natural habitat.　　 Otolith δ¹⁸O provides a method for fish population identification, migration and environment exploration of spawning ground. Otolith, a calcium carbonate mineral in bony fish, has both auditory and balancing functions and is known as the “recorder” of time and water environment. Microchemical analysis of otolith can reveal the life history and water environment changes of fish. Especially, the core area of otolith is formed at the early stage of incubation, and its microchemical composition can reflect the hydrochemical characteristics of the habitat and solve the problem of fish identification. The otolith δ¹⁸O is closely related to water temperature. Under normal salinity conditions, the δ¹⁸O change of 1‰ corresponds to the seawater temperature change of about 5℃. For example, the higher otolith δ¹⁸O values of Pacific bluefin tuna, which migrated across the Pacific from the Western Pacific (WPO) to the Eastern Pacific (EPO), reflected the cold-water temperatures encountered during migration. Similarly, the lower δ¹⁸O value in otolith core of Oncorhynchus keta, which migrated between freshwater and marine environments, indicated that Chinese salmon were also anadrotic spawning fish, returned to the ocean after hatching. Yuan et al.¹² analyzed the δ¹³C and δ¹⁸O values in the otolith of small yellow croaker in the Yellow Sea and the Bohai Sea, the population of southern Yellow Sea was subdivided into offshore and coastal populations for the first time, and no station crossing between the two populations was recorded. The overall successful rate of discrimination was recorded to be 82.6% by cluster analysis. Tatsuya et al.¹³ conducted otolith δ¹⁸O and microstructure analyses to investigate nursery habitat temperatures and early life growth rates. In conclusion, the otolith δ¹⁸O provides a method for fish population identification, migration, and environment exploration of spawning ground. In this study, the SHRIMP Ⅱ ion probe was used to determine the otolith δ¹⁸O of the naked carp in Qinghai Lake. Combined with the hydrochemical composition of lake water and river water in Qinghai Lake, the optimal migration and spawning sites and environmental conditions of the naked carp were investigated to provide effective information for further revealing its life history.

RESULTS The SHRIMP Ⅱ ion probe was used to determine the in-situ δ¹⁸O composition of the otolith microregion of five naked carp in Qinghai Lake. The in-situ oxygen isotope analysis of otolith was performed by SHRIMP Ⅱ ion probe at the School of Earth Sciences, Australian National University. Sectioned otoliths were prepared for SHRIMP analysis by casting them in epoxy resin, with NBS 18 and 19 reference calcites, to form a 35 mm diameter mount. After being documented by optical photomicroscopy, the samples were coated with high purity Al and transferred to the ANU SHRIMP Ⅱ for analysis using procedures based on those described in detail by Long et al.^21-22. In brief, the SHRIMP Ⅱ was operated in multi-collector, negative ion mode. A 15kV, ~3nA Cs⁺ primary ion beam was focused to a 25mm diameter spot on the Al-coated target, producing 200−250pA of secondary ¹⁶O⁻. ¹⁶O⁻ and ¹⁸O⁻ were measured simultaneously on Faraday cups using Keithley 642 electrometers. Each analysis consisted of a pre-burn of about 3 min to allow the secondary ion isotopic composition to stabilise, followed by 10 or 14 times (10s/time) estimates of the ¹⁸O/¹⁶O ratio. The accuracy was 0.1‰−0.2‰, and the standard deviation was about 0.3‰. The reference material was measured once every 5 samples completed during analysis. The spot starting position of the ion probe was the core of the otolith, along the longest growth axis to the edge. The spot diameter of the ion probe was about 25μm, and the spot depth was about 2−3μm. The spot test was shown in Fig.E.1A.　　(2) The δ¹⁸O composition along the growth axis of the otolith. The δ¹⁸O ratios along the growth axis of 5 naked carp otoliths were shown in Fig.E.1B. The mean values of otolith δ¹⁸O from No.1 to No.5 were 1.29‰, 1.79‰, 0.99‰, 0.93‰, −7.41‰ respectively, and their interval range were −4.88‰−3.46‰, −0.28‰−3.91‰, −1.43‰−2.94‰, −1.81‰−3.35‰, −9.36‰−5.21‰, respectively. The otolith δ¹⁸O in individual juvenile plaice ranged from 0.2‰ to 1.9‰, when they were reared at two temperatures (11℃ and 17℃), which indicated that the otolith δ¹⁸O ratios were significantly affected by water temperature, but not by feeding level, and there were no significant synergistic effects. The δ¹⁸O ratios of plaice are close to the higher δ¹⁸O ratios of samples 1 to 4 in this study, which is the result of the fact that naked carp lived in the lake with high salinity and low water temperature for part of the time. The Schizothorax kozlovi and the naked carp belong to the subfamily Schizostomus. The average otolith δ¹⁸O ratios of the Schizothorax kozlovi cultured in fresh water is −9.4‰, which is very close to the δ¹⁸O ratios of the naked carp No.5, indicating that sample No.5 has been living in fresh water. The gray bands are the incremental zones of otolith, and the interval zones between the incremental zones are discontinuous zones. The δ¹⁸O ratios of the naked carp in samples 1 to 5 show a trend of lower incremental zones and higher discontinuous zones, because the δ¹⁸O ratios of otolith is inversely correlated with water temperature. The incremental zones of otolith are formed in spring and summer, while the discontinuous zones in autumn and winter, and the water temperature in spring and summer is significantly higher than that in autumn and winter. The otolith δ¹⁸O ratios of the five samples were analyzed by one-way variance analysis. The δ¹⁸O ratios of the No.1, No.3 and No.4 naked carp were not significantly different (P>0.05), while the δ¹⁸O ratios of the No.2 and No.5 naked carp were significantly different from the other three samples (P<0.05). The δ¹⁸O ratios of Qinghai Lake ranged from 1.93‰ to 3.71‰, with an average value of 2.43‰, and that of surrounding rivers ranged from −8.04‰ to −6.20‰. The average δ¹⁸O ratios of Buha River, Shaliu River, Quanji River and Heima River were −6.15‰, −7.21‰, −5.98‰, and −7.31‰, respectively. This is because the δ¹⁸O ratios of different types of water bodies vary greatly, and the δ¹⁸O ratios of lake water are obviously higher than that of surface river water due to strong evaporation. Comparing the δ¹⁸O ratios of lake water and river water with that of 5 samples, it can be seen that the δ¹⁸O signals of lake water and river water are recorded in the otolith of No.1-No.4, indicating that they migrated between lake water and river water, however, the δ¹⁸O signal of river water was only recorded in the otolith of No.5, which indicated that the No.5 naked carp had been living in river water, that is, “sitting water fish”.　　(3) Composition of δ¹⁸O in otolith cores of the naked carp. The average δ¹⁸O ratios of otoliths from No.1 to No.5 were 1.51‰, 0.20‰, −0.22‰, −2.38‰, and −9.14‰, respectively, and their interval range were 1.43‰−1.59‰, 0.28‰−0.7‰, 1.02‰−0.64‰, 4.88‰−0.74‰, 9.36‰−8.96‰. The average δ¹⁸O ratios of otolith cores were compared with the δ¹⁸O composition of lakes and rivers, as shown in Fig.E.1C. Since the core area of otolith was formed in the early incubation period of naked carp, the water environment at birth was recorded. From the distribution, the δ¹⁸O ratios of the core area of naked carp No.1-No.3 were closer to the lake water, which indicated that their spawning place was near the estuary. The water body was mixed with lake and river water, and having lower water temperature, which indicated the migration and spawning may have started in late May. The δ¹⁸O ratios of the core area of the No.4 naked carp was closer to the river water, which indicated that the spawning area was far away from the lake and the water temperature was higher, while the No.5 naked carp lived in the river all the time and the water temperature was higher when spawning. In conclusion, the habits of the naked carp are not fixed. Some breed in estuaries, some in upstream rivers, and some migrate between rivers and lakes to spawn, and a small part of the naked carp breed in rivers, which may be related to the local rivers’ cutout in autumn and winter. Some fish that migrate to the upper reaches of rivers to spawn have stayed in rivers for breeding due to the cutout of rivers. Similarly, the otolith cores of Pacific herring born in the Strait of Georgia have lower δ¹⁸O ratios (−8.2‰ to −2.0‰), while the otolith cores of the southern Puget Sound samples have higher δ¹⁸O ratios (−3.9‰ to −0.9‰).

DISSCUSION (1) The relationship between δ¹⁸O ratios of otolith and water temperature. The otolith of the naked carp can record not only the age, but also the water environment and the life history of the fish. The otolith δ¹⁸O ratios of fish is determined by the δ¹⁸O ratios of water body and the water temperature. The δ¹⁸O ratios of the water body are related to the ratio of precipitation to evaporation, runoff and precipitation, while there is an inverse correlation between temperature and otolith δ¹⁸O ratios. The living environment of fish can be inferred from the δ¹⁸O ratios value on the otolith of fish, and the isotopic balance between the surrounding water environment and aragonite minerals is reached or almost reached, therefore, the fractionation of ¹⁶O/¹⁸O isotopes is closely related to the temperature of the water body, generally showing a negative correlation. For marine fish, the variation of δ¹⁸O ratios at 1‰ corresponds to the variation of sea temperature at about 5℃ under normal salinity conditions. The relationship between temperature variation and otolith δ¹⁸O ratios of freshwater fish at 1‰ has not been established.　　(2) Experimental research about temperature control on otolith δ¹⁸O ratios. Water temperature, water flow, water depth, riverbed quality and light are the most important environmental parameters in fish breeding. In general, the reproductive activities of fish are not determined by a single environmental factor, and their specific reproductive needs are often multiple environmental factors. When the water temperature of the freshwater tributaries around Qinghai Lake reaches the threshold of 6−16℃, the naked carp will start to migrate, on the one hand, due to the different composition of δ¹⁸O ratios in different spawning sites, and on the other hand, due to the different spawning time, the difference of water temperature is recorded in the otolith δ¹⁸O.The contribution of the δ¹⁸O composition and the change of water temperature to the δ¹⁸O ratios in the otolith of the naked carp cannot be quantified, and needs to be further verified by the artificial control experiment. For example, Nakamura et al.³¹ cultured mackerel larvae at six different temperatures (16.3℃, 17.6℃, 18.3℃, 20.0℃, 24.0℃ and 26.5℃) and determined the δ¹⁸O ratios, the linear relationship between the water temperature and the otolith δ¹⁸O of mackerel was determined. Willmes et al.³² reared male oversea fish for 360 days at different salinity (8.75‰, 5.28‰ and 4.06‰) and water temperature (16.4℃, 16.7℃, 18.7℃ and 20.5℃), and then reconstructed the thermal life history of the fish. Wang et al.³³ evaluated whether cooking behavior would lead to further isotopic fractionation of carbonate in otolith and showed that the otolith δ¹⁸O ratios were highly consistent between cooked and uncooked, which indicated the cooking process has no or little effect on the otolith isotope value.　　(3) Fossil otolith is a good carrier to study the paleoclimate environment. Otoliths are well preserved in sediment, so the fossil otoliths are often used in paleoclimate, palaeoecology and palaeogeography. For example, Wurster et al.³⁴ analyzed high-resolution δ¹⁸O ratios of Aplodinotus grunniens from the Easterman archeological site to infer climate change in the eastern continental United States during the Holocene. Andrus et al.³⁵ used the otolith δ¹⁸O ratios of a well-preserved Peruvian catfish fossil to calculate the sea surface temperature of the Peruvian Sea in the Middle Holocene. Surge et al.³⁶ analyzed the δ¹⁸O ratios of modern and circa 2nd−3rd century AD fossil otoliths of the Mexican sea catfish, and the recovered winter temperatures were very similar to those of modern times. Long et al.²¹ analysed the δ¹⁸O ratios of otolith in the sediment of Mungo Lake, Australia, and the variation of the δ¹⁸O ratios indicates an early flood in the lake, a subsequent increase in the δ¹⁸O by 4‰ indicating increasing evaporation of the lake. At the same time, because there is no “Reservoir effect”, the otolith preserved in the sediment is also a good material to measure radioactive carbon. Therefore, fossil otoliths preserved in sediments are not only one of the ideal materials to determine the age of sediments, but also a good carrier to study the paleoclimate environment.　　(4) Relationship between δ¹⁸O ratios of otolith and fish breeding population. The naked carp in Qinghai Lake is one of the second-class national protected animals in China. Since 1982, Qinghai Province has implemented the closure of the lake to breed fish, covering Qinghai Lake and all the rivers entering the lake. Currently, the sixth closure period in 10 years will end on December 31, 2030. At the same time, the original species preservation bases and artificial breeding and releasing stations have been established around Qinghai Lake. Since 2002, artificial breeding and releasing work has been carried out, and 105 million large-size naked carp fry have been released into the Shaliu and Quanji rivers. Therefore, there should be two populations of natural reproduction and artificial reproduction in Qinghai Lake, and it is difficult to distinguish the two populations only by the otolith δ¹⁸O ratios.　　(5) In order to identify fish populations more accurately and understand fish life history, many scholars combine the δ¹⁸O and δ¹³C ratios of otolith for analysis, because the change of otolith δ¹³C ratios can record the maturity of fish and the change in the food chain. For example, Gao et al.³⁸ analyzed the δ¹⁸O and δ¹³C ratios of the otolith core and the second summer season round of Pacific herring in Pikew Bay, Washington, and concluded that there are two types of herring. Ashford et al.³⁹ tested the δ¹⁸O and δ¹³C ratios of cod otolith located on the Patagonia continental shelf and Patagonia in southern Georgia, and found that the δ¹⁸O ratios reflected the sea water temperature in the surrounding environment where the fish lived, so that the different population could be identified. The source of metabolic carbon and the δ¹³C ratios of dissolved inorganic carbon (DIC) in seawater affect the δ¹³C ratios of otolith. For the first time, Jiang Tao et al.⁴⁰ analyzed otolith stable isotope ratios of δ¹⁸O and δ¹³C using an isotope ratio mass spectrometry on Coilia nasus juveniles collected from the Changjiang River Estuary. The different environments with different water temperature and food organism composition might be experienced by the two groups of C. nasus fish. He Yongfeng et al.²⁵ took Sichuan schizostomus of different age groups as the research object to explore the relationship between stable isotopes and environment. The results showed that δ¹⁸O and δ¹³C ratios of Schizothorax kozlovi at age 1t were not significantly correlated with otolith mass, but both were significantly different between lapillus and asteriscus. There was no significant difference of lapillus δ¹⁸O and δ¹³C between the sexes, but with significant difference among different ages. The correlation of δ¹⁸O with δ¹³C was an effective method in identification of different culture stocks of S. kozlovi, which indicates that the isotopic signatures of otolith could be used as a method to identify freshwater fish stocks.