**1. Introduction**

The quantitative and qualitative state of water resources of a watershed are formed by a variety of drivers that interact in a complex and often indirect way [1]. Climate elements (precipitation; relative humidity; wind speed and direction; solar radiation and temperature that also controls evaporation/evapotranspiration and snow melt and their temporal and spatial distribution) and the biogeophysical characteristics of a catchment (topography, land—vegetation cover, geological structure, soil coverage) are fundamental determinants of regional hydrology [2,3]. The conceptual model describing the interactions among the abovementioned drivers is the hydrological cycle that links the exchange, storage and movement of water among the biosphere, atmosphere, cryosphere, lithosphere, anthroposphere, and hydrosphere [4], while the quantification of the relationships among the components of the hydrological cycle at a given location constitutes the water balance [5].

All characteristics of the catchment (climate elements and biogeophysical characteristics) are factors that can be largely affected by anthropogenic activities and pressures [3]. Humanly imposed climate

change due to increased emissions of greenhouse gases and dust from anthropogenically-disturbed soils [6], is expected to significantly increase freshwater-related risks such as modification of the hydrological regime, floods and droughts, and to a ffect water cycle components [7,8]. Climate change is projected to reduce renewable surface water and groundwater resources significantly in most dry subtropical regions and is likely to increase the frequency of meteorological droughts (less rainfall) and agricultural droughts (less soil moisture) in presently dry regions. Additionally, projections imply variations in the frequency of floods and negative impacts on freshwater ecosystems by changing streamflow and water quality [4,7]. Regarding anthropogenic interventions in catchment's physical characteristics (for example alteration of the land surface soil moisture, albedo and roughness [9]), land cover changes due to livestock grazing, agriculture, timber harvest, deforestation, and urbanization can reduce retention of water in watersheds and lead to an increase of the size and frequency of floods and to the reduction of baseflow levels [10]. Dam constructions and diversion, canalization, snagging and dredging of rivers, streams and drainage ditches, and groundwater overexploitation, disrupt the dynamic equilibrium between the movement of water and sediment that exists in rivers [10]. Based on recent studies, direct human impacts on the terrestrial water cycle are in some large river basins of the same order of magnitude, or even larger than climate change [11,12]. Especially land cover change alters annual global runo ff to a similar or greater extent than other major drivers [13], while land use change contribution in regional runo ff values in tropical regions is larger than that of climate change [14], especially in the case of smaller catchments [15].

Worldwide studies support the impacts of land cover changes, mainly deforestation and urbanization, on the hydrometeorological factors, leading primary to river discharge increase [16–23] and generally to an increase of eco-environmental vulnerability of the watersheds [24,25]. In Greece, studies confirm the impact of land cover change and deforestation in river discharge. For example, a study conducted in Pinios river basin proved that expanding the agricultural land over forest by 20%, a mean monthly increase in the river discharge of up to 3%, can be observed from October to April and a respective reduction from May to September, reaching a maximum of 6% in July [26]. Moreover, human interference in streams crossing urban or suburban areas raise the vulnerability to flash floods. For example, the hydrometeorological analysis of a fatal flash flood event which occurred on 15 November 2017 in the suburban area of Mandra, western Attica, Greece resulting in extensive damages and 24 fatalities, showcased heavy storm-induced run-o ff water in combination with human pressures on streams as the reason for the flood [27].

Regarding the impact of land cover change to evapotranspiration, it has been reported that mean annual evapotranspiration can be up to 39% lower in agricultural ecosystems than in natural ecosystems in Brazil [20]. A recent study concerning the whole of China showed that the average annual land surface evapotranspiration decreased at a rate of −0.6 mm/yr from 2001 to 2013, attributed partly to land use and land cover changes of forests to other land types [28]. In Greece, a study in a small catchment showed that 16% increase of agricultural land against wetland and forest area led to a 6% increase of evapotranspiration and 10% increase of the water deficit in the soil [29].

Given the uncertainty of future land cover changes due to socio-economic driving forces and local development policies applied, a scenario-based modeling framework can be beneficial in supporting the analysis of potential land cover changes, so as to mitigate potentially negative future impacts on a basin's water resources. In order to investigate the e ffect of anthropogenic land cover changes to the hydrological cycle components and the main hydrometeorological factors of a regional agricultural watershed in Central Greece (Spercheios river basin) of grea<sup>t</sup> ecological value, three (3) land cover case studies were adopted, based on the land cover distribution documented in the following years: in 1960 (hereafter LC1960; baseline), in 1990 (hereafter LC1990; mid-period), and in 2018 (hereafter LC2018; current state). The modeling tool used was the physically-based hydrological model (MIKE SHE), while the high-resolution gridded observational daily meteorological dataset of Europe named E-OBS [27] from the EU-FP6 project UERRA [30] and the Copernicus Climate Change Service [31] was also employed to drive the model. Since the E-OBS gridded dataset had not been used before in similar studies in Greece, the statistical evaluation of its efficiency was considered to be obligatory before performing any further analysis. Finally, statistical tests and trend analysis were performed on the simulated time series of each land cover case study examined.

The main objective of the present study was the better understanding of the system's response and the basin's water resources to possible future land cover changes, while the main research questions intended to be addressed are: (a) which are the interrelationships among land cover and the main hydrometeorological factors' (precipitation, air temperature, discharge, and actual evapotranspiration) variations, (b) how land cover changes affect the trend magnitude of the main hydrometeorological factors, and (c) which are the hydrometeorological-related hazards associated with land cover changes in the study area?

#### **2. Materials and Methods**
