1 Objective

The aim of collecting these data is to have a long series of a proxy of salinity in an estuary of great economic interest for its commercial bivalve mollusc populations. Although there are data loggers in the estuary, they only measure salinity in the water column at a tidally variable distance from the bottom. However, salinity affecting bivalve mollusc populations is that recorded no more than 1 m above the bottom and at high tide [1]. The values included in the data series are the results of an ocean-meteorological model that considers the runoff from the Ulla river basin flowing into the estuary, tidal amplitude, and wind direction and strength. Data from 2002 to 2003 were used to build the model and data from 2004 to 2006 for its validation [1]. The rest of the data have never been used in any research paper. However, some point data between 1977 and 2009 were used for the definition of different categories of mortality events for commercial bivalve stocks [1]. Data after 2002 have been used in internal management reports for the cockle fishery in the estuary.

Although the data resulting from this model have been validated with those observed in situ, the river Ulla is regulated by a hydroelectric power station that can alter its flow. However, rather than a limitation, this circumstance could be an opportunity to use the data as a control of possible deviations of the salinity variation due to ocean-meteorological conditions, as an effect of the regulation of the river flow by the dam. Another possible use of this long dataset is the study of variations in the frequency and intensity of low salinity episodes that could lead to mortality events of different intensity of molluscs of commercial interest in relation to climate change.

2 Data description

The dataset consists of two series. The first one (Data file 1 in Table 1 [2]) is composed of daily salinity data for 1 m above the seabed during high tide in the estuary of the Ulla river (Ria de Arousa, Galicia, Spain) over 63 years. The data were originated by a model that considers the runoff in the river basin [3], the tidal amplitude, and the wind direction and strength. Details of the methodology, construction and validation of the model can be found in Parada et al. [1]. The file consists of two columns: the first contains the dates and the second the daily salinity values. Missing values have been identified as “-999”.

Table 1 Overview of data files/data sets

The second series (Data file 2 in Table 1 [4]) records the number of critical salinity events occurring in the hydrological years 1961 to 2022 (62 years in this case). A hydrological year is defined as the one starting in October and ending in September of the following year. The series includes four of the possible critical episodes defined by Parada et al. [1]: Salinity (S) below 10 for 1 day (S10-1d); S < 30 for 18 consecutive days (S30-18d); S < 5 for 1 day or more (S5-1d) and S < 30 for 19 consecutive days or more (S30-19d). The first two types of events can cause moderate mortalities of commercial bivalve molluscs and the second severe mortalities [1]. Moderate mortalities are defined as events where mortalities of less than 50% of the cockle (Cerastodema edule, L. 1758) and up to 15% of the clams Ruditapes decussatus (L., 1758) and R. philippinarum (A. Adams & Reeve, 1850) are recorded in the estuary. Severe mortalities are defined as events in which mortalities of 50% or more of the population of C. edule and the clam Venerupis corrugata (Gmelin, 1791), and 15% or more of R. decussatus and R. phillipinarum are recorded [1]. The file is composed of 5 columns: the first column contains the hydrological year and the following columns contain the number of episodes recorded for each of the four categories (S10-1d, S30-18d, S5-1d and S30-19d).

3 Limitations

  • The data in data file 1 [2] are from the output of an ocean-meteorological model, not direct salinity measurements and are considered representative to the last 3 km of the Ulla estuary.

  • The model does not consider the existence of a dam in the river that flows into the estuary. However, no deviations have been observed when the model was validated with real measurements.

  • There are 373 missing data in the data file 1 [2] due to the punctual lack of meteorological data feeding the model. The missing data have been replaced by the value “-999” and represent 1.6% of the 23017 data of the series.