LAGOON WATER-LEVEL OSCILLATIONS DRIVEN BY RAINFALL AND WAVE CLIMATE

11 Barrier breaching and subsequent inlet formation represent critical processes that ensure the 12 temporary or permanent connection and transference of water, nutrients, or living organisms between a 13 lagoon and the open sea. Here, we investigate the conditions inducing natural barrier breaching through 14 a 34 months monitoring program of water-level oscillations within a shallow lagoon and the adjacent 15 nearshore, at the Northern coast of the Iberian Peninsula, Louro lagoon. Seven natural openings were 16 identified during the three monitored wet seasons (Wet1, Wet2 and Wet3), four in the Wet1, two in the 17 Wet2 and 1 in the Wet3. Identified openings were grouped in three types depending on the observed 18 relation between the lagoon water-level (Lwl), the berm height (Bh) and the water-level at the beach (Bwl): 19 (i) openings by lagoon outflow, which include those characterized by Lwl higher than the Bh and lower 20 Bwl; (ii) openings by wave overwash, including those induced by Bwl higher than the Bh, and (iii) mixed 21 openings, which result from a combination of the two previous conditions. We have found that the Lwl 22 is modulated by the rainfall regime (Rf) and can be explained by the accumulated precipitation while Bh 23 and Bwl depend on the wave climate and tidal level and can be estimated applying runup equations. The 24 inlet lifespan was found to be regulated by the wave climate and rainfall regime; in particular barrier 25 sealing was associated with a sudden increase in wave period and reduction in precipitation. This work 26 proves that the natural openings could be predicted successfully with support to medium term water27 level monitoring programs, which in turn may significantly contribute to strategic decision making for 28 management and conservation purposes. 29


INTRODUCTION 33
Coastal barrier breaching (inlet formation) is a complex morphodynamic process that enables free 34 water exchange between the lagoon and the open sea. Processes performing at both the seaward and the 35 bay side of a barrier may induced barrier breaching (Boothroyd, 1985;Fagherazzi and Priestas, 2012; 36 the North Atlantic Oscillation -NAO- (Goodess and Jones, 2002;Osborn et al., 1999). The rainfall in 115 the study area is highly variable, with an average rainfall ranging from 600-700 mm in winter to less 116 than 100 mm in summer. The average annual rainfall is close to 1700 mm (Martínez Cortizas and Pérez 117 Alberti, 1999). Previous works suggested that precipitation in this region is strongly modulated by the 118 NAO, with more humid conditions during winters corresponding to low and negative NAO index values 119 (Rodriguez-Fonseca and de Castro, 2002; Trigo et al., 2008). 120

MATERIAL AND METHODS 121
Two water-level loggers were deployed at the study site to monitor water-level oscillations ( Figure  122 1, Table 1). Available topographic data (Table 1)  The DTM were constructed based on the Geodetic Reference System ED50 and were represented on the 128 UTM mapping projection (UTM zone 29N). All heights were referred to MSLA. 129 Metocean data were also examined and jointly analyzed with in-situ water levels: (i) meteorological 130 data (i.e. rainfall and evaporation parameters) from a near coastal meteorological station -Corrubedo 131 station ( Figure 1) -(downloaded from http://www.meteogalicia.gal/web/index.action), and (ii) wave 132 climate data from one node of the SIMAR dataset (courtesy of Puertos del Estado), which includes 133 hindcast winds, sea-level and waves starting in 1958 ( Figure 1, Table 1). 134 Lagoon water-level monitoring 136 One logger was located in the lagoon to register lagoon water-levels (Lwl; Figure 1). Water 137 oscillations were recorded at 5 min time intervals, for 34 months (Table 1) with Water-level logger 138 models Seabird SBE 39 and AQUALogger 520 PT. The elevation of the logger was measured, once 139 deployed, using a Trimble DGPS-RTK. Observations were therefore referred to the MSLA by 140 referencing to the corresponding water logger elevation. Water-level measurements were corrected for 141 variations in barometric pressure using the data downloaded from the closest meteorological station 142 ( Figure 1). 143 Corrected data were used to determine the timing of lagoon natural opening, the duration of the active 144 inlet phase, inlet sealing and the associated water-level thresholds. In addition, these data allowed us to 145 identify the parameters that can characterize the lagoon openings: (i) the plateau phase, which 146 corresponds to the time (in hours) between the moment when the lagoon reaches its highest water-level 147 and the breach, and (ii) the water-head (or hydraulic head) difference (in meters), which was calculated 148 as the difference between the lagoon water-level and the water-level in the nearshore at the opening. 149 Rainfall regime impact on lagoon water-level 150 The relation between Lwl and the rainfall was evaluated using the accumulated rainfall (Rf) for the 151 periods when the lagoon was closed. First the periods with more intense and frequent rainfalls or wet 152 seasons were identified, including also the periods during which the lagoon opens. In addition, the water 153 storage capacity of the lagoon was obtained using the bathymetry to translate lagoon water-levels into 154 water-volumes (Lv). Therefore, the relation between Lv and Rf was obtained, which can be applied for any situation knowing the accumulated rainfall. In addition, the obtained expression can only be applied 156 after a certain level in the lagoon is reached in order to allow direct comparison between events (Lmwl). 157 To select the latter, we have imposed criteria to normalize all the data that is defined by local sea water 158 levels: 159 where MW is sea mean water-level and MHW is the high sea mean water-level 161 3.2. Nearshore water-levels 162 Tidal regime 163 The tidal regime was monitored using a logger located at the beach nearshore ( Figure 1). Water 164 oscillations at the nearshore were recorded at 5 min time intervals, for 28 months (Table 1). We used 165 the same models of water-level loggers as previously described for the Lwl monitoring. The same 166 topographic and barometric corrections used for the lagoon water-level record were applied to these 167 data. The corrected record was analyzed using the script World Tides To ensure the correct application of equation 2, the simulated nearshore waves were reversed shoaled 201 to deep-water using the linear wave theory, and assuming a shore-normal approach and the unrefracted 202 wave height and period as suggested by Stockdon et al. (2006). 203 A similar approach was used to estimate a range in the elevation of the sandy barrier (Bhmin -Bhmax) 204 by assuming that Bwl during antecedent spring-high tides is a proxy for beach berm elevation. The latter 205 is in turn assumed as representative of the barrier dimensions at the area where the barrier breaches, 206 which in turn lacks any dune building.   Table 2. Detailed information of water-levels into the lagoon, rainfall, tides and waves for opening events. Initial water-levels in the lagoon are established as the minimal water-level before a breach (Lmwl). Beach berm values (minimal and maximum) are calculated by adding to the height of the spring-high tides values the corresponding runup values before a barrier beach. The plateau phase corresponds to the time (in hours) between the highest water-level reached by in the lagoon and the breach. The barrier recovery time is referred to the time between the closing of the inlet. The next opening is given in days. The water-head (or hydraulic head) difference (in meters) was calculated as the difference between the water-level in the lagoon and in the barrier. Accumulated rainfall corresponds to the rainfall drop lapsing between the moments that the lagoon reaches the Lmwl and a barrier breaching event. The event marked as * corresponds with a maximum water-level not leading to an opening. Different font colors in the events correspond to different wet seasons. only one event during Wet3. The first opening event of each wet season was characterized by Lwl above 4 m 234 (Table 2) while consecutive openings within a same season were below 4 m, which means that the time interval 235 for barrier recovery would be relatively short; openings 2, 4 and 6 ( Table 2). 236 Figure 3A shows the relationship between Lwl and the corresponding Lv for Lwl higher than 2.91 m (Lmwl 237 obtained using equation 1). The relation (with a r 2 value of 0.99) between these variables was: 238 (4) = 259641 − 643018 239 Figure 3B shows Lv versus Rf for each opening event between the moments in which the imposed criteria 240 is attained (Lwl=Lmwl) and the breaching moment. The relation (with a r 2 value of 0.75) between these variables 241 can be described with by the following regression: 242  To validate this relation, we have used the values of rainfall in our study area and the recorded lagoon 247 water-levels. Figure 4 represents the Lwl, recorded and predicted using equation 6, at the barrier breach. The 248 predicted values are close to the recorded values (less than 0.3 m of difference) with the exception of the events 249 2 and 3, having a difference of 0.5 and 0.8 m respectively. 250   (Figure 2A). Identified openings occurred at 256 spring tides or close to spring tides with the exception of opening 7, which took place close to neap tides. Four 257 of the recorded openings occurred close to the high tide (openings 1, 3, 4 and 6) while the other three (opening 258 2, 5 and 7) happened close to the low tide (Table 2). 259 propagated waves with SWAN model at points 1 and 2 (Figure 1). The results show that higher wave heights 261 came from SW, suggesting the occurrence of storms recorded during the wet seasons. Waves from SW impact 262 the beach directly while waves from westerly and northerly directions are transformed before reaching the beach due to wave refraction. The effect of refraction is not linear and becomes higher as the offshore waves 264 approach NW, reaching a maximum difference of 50º between offshore and local wave direction (e.g. 343º -265 NNW-transformed to 290º-WNW-). The wave height and period were reduced by 30% and 20% respectively. 266 Differences between the nearshore points 1 and 2 were only observed on waves above 3m, suggesting a higher 267 effect as the waves enter the northern-end of the bay. 268 The minimum and maximum values estimated for barrier elevation (Bhmin and Bhmax) before the openings 269 are presented in Table 2, ranging from 3.7 to 5.3 m. The latter were associated with local waves arriving 270 parallel to the beach or with low angles (≈225-230º) and high periods (>8 s), promoting onshore sediment 271 transport. 272 In the same way, Bwl at the openings were calculated and are presented in Table 2. The estimated values of 273 Bwl for four of the identified openings were lower than the recorded Lwl (openings 2, 3, 5 and 7), resulting in a 274 positive water-head difference. Alternatively, the other three cases estimated Bwl values were higher than Lwl 275 (openings1, 4 and 6), producing a negative water-head difference ( Table 2). 276

Processes and data integration 277
Barrier breaching was tentatively parameterized using the relation between the Lwl (lagoon water-level); Bh 278 (Barrier height) and Bwl (barrier water-level). Wave climate, rainfall and the tidal range modulate the selected 279 parameters. Indeed, the water level in the lagoon can be predicted using the accumulated rainfall, while the 280 elevation of the berm can be estimated using the local wave climate and nearshore water level. In addition, the 281 relation between these parameters determines the mechanism that will induce barrier breaching and could be 282 used to predict the timing, the direction of the lagoon openings and, therefore, the type of opening. Three types 283 of openings have been identified: (i) Lw-type or breaching triggered from the lagoon, (ii) Sw-type or breaching 284 triggered from the sea and (iii) Mx-type or mixed lagoon-sea opening. 285  (Table 2). This type was also associated with a strong barrier, 290 with more than 48 days to recover from a previous opening (Table 2). In addition, the water-head difference (Table 2), always showed positive values greater than 1 m, generating a gradient 292 between the lagoon and the sea side. Lw-types were preferably happening with low waves heights 293 and spring tide, only the event 7 occurred with high height waves but at low and close to neap 294 tides (Table 2). We observed that for all these cases the relation Lwl ≥ Bhmin > Bwl was filled. 295

(ii)
Sw-type. Openings 4 and 6 ( Figure 6, event 4) were classified as Sw-type. In both cases, the 297 opening occurred shortly after beach berm reconstruction, following barrier breaching within the 298 same wet period. The water-level inside the lagoon and the accumulated rainfall was lower than 299 for the Lw-type, with values circa 3.5 m of water-level and circa 80 l/m 2 for rainfall (Table 2). 300 Estimated beach berm elevations before the opening were similar or lower than the values 301 obtained for the Lw-type (see Table 2). However, the barrier recovery time in these cases was less 302 than 20 days (Table 2), and the SW waves reached the beach at an oblique angle to the shoreline. 303 These waves were previously documented as responsible for the beach-face erosion in the study 304 area, provoking the lowering and narrowing of the barrier ). Water-head 305 differences were negative and greater than 1 m for these cases, generating an inverse gradient 306 between the lagoon and the ocean. During these events, the plateau phase was not present. Sw 307 type events developed during spring tides, coincident with high tides and high SW waves 308 promoting high runup values (see Figure 2 and Table 2). For these cases the observed relation 309 between the three variables was Bwl ≥ Bhmax > Lwl. 310

(iii)
Mx-type. This type represents the openings 1 and 2 (Figure 7, event 2) that could not be easily 312 grouped within the Lw-and Sw-types. The water-level inside the lagoon was relatively high 313 (around 4 m, Table 2). Yet, the water-head difference in this type was positive or negative but 314 lower than 1 m. Moreover, like in the Sw type, the days before the opening were characterized by 315 high SW waves, with high erosion potentials to erode the beach berm, inducing barrier breaching.
Under such conditions, it is expected that the weak barrier would not be able to store high water-317 volumes in the lagoon, maintaining the plateau phase only for less than 16 hours. For that type of 318 opening, the relation between the variables was Lwl ≈ Bhmin < or > Bwl. 319  (Table 2). In turn, inlet longevity at Louro seems to be regulated by the rainfall regime and wave climate 326 after the opening. The identified closure events were coincident with a cessation of rain, which would explain 327 the reduction of the water input, and the incidence of constructive waves characterized by higher period values. 328 The latter would promote the onshore sediment transport and the development of high berms with the