Factors controlling erosion-deposition phenomena related to 1 lahars at Volcán de Colima , Mexico 2

One of the most common phenomenon at Volcán de Colima is the annual development of lahars that 6 runs mainly through the southern ravines of the edifice. Since 2011 the study and the monitoring of these flows 7 and of the associated rainfall has been achieved by means of an instrumented station located in Montegrande 8 ravine, together with the systematic surveying of cross topographic profiles of the main channel. From these, 9 we present the comparison of the morphological changes experimented by this ravine during the 2013, 2014, 10 and 2015 rainy seasons. A total of 11 lahars occurred during this period of time, and their erosion/deposition 11 effects were quantified by means of the cross-section areas determined from the profiles taken at the beginning 12 and at the end of the rainy seasons and before and after the major lahar event of 11 June 2013. From the data 13 compiled in these surveys, we identified two main zones: i) an erosive zone between 2100–1950 m a.s.l., 8° in 14 slope, with progressive channel bed deepening and/or widening, and with an annual erosional rate of 10.3% 15 due mainly to the narrowness of the channel and its high slope angle and, ii) an erosive-depositional zone, 16 between 1900–1700 m a.s.l., (~8% erosion and ~16% deposition), due to a wider channel and that decreases in 17 slope angle (4°). These observations were confirmed by simulating with the FLO-2D code a flow with a 18 hydrograph similar to the 11 June 2013 lahar, the largest event observed during the investigated period. Based 19 on these observations, the major factors controlling the erosion/deposition rates are the channel-bed slope, the 20 cross section width and the joint effect of sediment availability and accumulated rainfall. On the distal reach of 21 the ravine, the erosion or deposition processes tends to be promoted preferentially one over the other depending 22 mostly upon the width of the active channel. Only for extraordinary rainfall events, lahars are mostly erosive 23 all along the ravine up to the distal fan where deposition take place. Finally, by comparing rainfalls associated 24 to lahars originated after the last main eruptive episode that occurred in 2004-2005, we observed that higher 25 accumulated rainfalls were needed to trigger lahars in 2013 and 2014 seasons, which point to a progressive 26 stabilization of the volcano slope during a post eruptive period. These results can be used as a tool to foresee 27 the effects of future laharic events in Volcán de Colima and to improve the input parameters for the modelling 28 of these flows, in order to better constrain the hazard zonation for lahars in this volcano. 29


52
As observed by Lavigne (2004), on these types of volcanoes, the efficiency of erosion is a consequence of rain-53 triggered lahars that develops during the rainy season, converting the estimation of sediment yield and erosion 54 rate a very difficult task to achieve (Lavigne, 2004;Procter et al., 2010;Pierson et al., 2011;Thouret et al., 55 2014). However, several studies have been realized in order to analyze the erosional and depositional processes 56 of active channels on volcanic environments along with the factors that controls them (Major et al., 2000; 57 Lavigne, 2004;Berger et al., 2011;Pierson et al., 2011;Starheim et al., 2013;Thouret et al., 2014). Based on 58 these studies, main factors affecting erosion and depositional rates are: the amount of rainfalls and volume of 59 sediments available, the hydrologic characteristics of the stream-bottom deposits, flow depth, bed-slope 60 gradient and the morphology of the channel (Mizuyama and Kobashi, 1996;Fagents and Baloga, 2006;Berger 61 et al., 2011;Okano et al., 2012;Thouret et al., 2014). Moreover, a number of studies address the issue of 62 identifying governing factors of debris flow entrainment and most of them have been carried out in laboratory 63 (Mangeney et al., 2010;Iverson et al., 2011) or in mountain environments (e.g. Chen et al. 2005;Hungr et al.

64
2005; Guthrie et al., 2010;Berger et al., 2011;McCoy et al., 2012;Abancó and Hürlimann, 2014;Theule et al. the field campaign of 30 July 2013, two lahars formed; after this and until the next field campaign on 11 October, 131 just one lahar took place (see Table I). During the 2014 season, 7 lahars developed until the field campaign on 132 12 September; and between these surveys and the survey taken on 18 March 2015, the only lahar developed 133 was on 17 March 2015 (Table I). Thus, the morphological changes described in the following sections are 134 related to the beginning and at the end of these seasons.

135
To quantify the erosion/deposition rates, for each section, a discrete areal value was computed and used to 136 calculate it (e.g. Fig. 3a and   3) There will be erosion (E), if between Date t1 and Date t2, A3 > A1 and A4 < A2 148 4) There will be deposition (D), if between Date t1 and Date t2, A3 < A1 and A4 > A2

150
From these remarks, the areal values AT, A1, A2, A3 and A4, along with the rates E and D, were estimated for   (Table I). It was the 156 first lahar of the 2013 season, and has been one of the biggest events recorded in the ravine since 2011. It 157 consisted of a flow that lasted approximately 3 hours, and presented several FEBs, followed by DSs. The total 158 rainfall associated to this lahar was 117 mm accumulated in ~3.5 hrs with a maximum peak intensity of 131 159 mm/hr (Table I). For this lahar, we took profiles the day before and the day after (Figs. 3c, d, and 4) the event, 160 which allowed us to outline quantitatively the E and D rates (Fig. 7f), and served as an example of the effects  that filled up the channel bed by 1 m, at least, but eroded the right later terrace. In general, along the ravine, 167 erosion was predominant in the channel bed, than in the walls; however in some portions of the ravine, the 168 erosion tends to wash away preferentially one side of the channel than the other, developing a lateral migration 169 of the channel axis (e.g. Fig. 4a and c).

170
The width of the active channel also varied after the 11 June 2013 lahar, becoming ~2 m wider especially on Besides the MPE of 11 June 2013, three more lahars were developed during the 2013 rainy season (Table I),

174
for which additional observations are here provided for that year. In 2014, more than 5 flows were observed, 175 classified and analyzed, and the annual balance is here presented based on a survey before and after the season.

176
Finally, for 2015 only data gathered after the first lahar of the season is here described (Figs. 5, 6, and Table I).
177 Figure 5 shows the profiles of the channel bed of Montegrande ravine, from the monitoring site (at ~2050 m 178 a.s.l., Fig. 5a), towards the mouth of the ravine (at ~1750 m a.s.l., Fig. 5g). In all profiles, the same color 179 corresponds with the same year, and the gradient of its color corresponds to different field campaigns. The          widening of this part of the channel was preferentially towards the right bank of the bed axis (Fig. 6d).

249
Finally, on checkpoint MG_05 (Fig. 6e), the level of erosion during 2013 season was more evident with a value 250 of ~8% that correspond with the widening of the channel from ~8 m to ~16 m (Fig. 6e), but with a deposit in  In general, at checkpoints MG_01 and MG_02 ( Fig. 6f and g), the rate of erosion was higher than in the down 257 flow checkpoints (MG_03, MG_04, and MG_05, Figs. 6h, i, and j, respectively), but not that high in comparison 258 to the 2013 season (Figs. 6a to e, and 7f).

259
On checkpoint MG_01 (Fig. 6f), the erosion was mainly focused on the walls of the terraces that flanks the 260 channel axis, eroding ~8 % of the material, diminishing its height by 2 m, but conserving the active channel 261 width in ~2.5 m with a final deposit of ~1 m in thickness (Fig. 6f). On the other hand, at checkpoint MG_02, 262 both erosion/deposition are present, (Fig. 6g), with a final deposition rate of 0.7% (Fig. 7f) (Fig. 6i). Finally, at checkpoint MG_05 (Fig. 6j), a morphology similar to the 271 MG_03 site can be observed, where new material deposited preferentially on the right side of the channel wall 272 (D = 9.6%), leaving vertical steps and narrowing the channel from ~16 m to ~14 m (Fig. 6j).

286
The 11 June 2013 lahar represents an exceptional event that was erosive almost all along the ravine. This lahar 287 was associated with an extraordinary rainfall event (117 mm of accumulated rain, Table I) that developed a 288 large and highly erosive flow. Similar exceptional events have been previously observed in the same ravine 289 during the Jova hurricane, which also developed a high-magnitude lahar that deeply eroded the channel (Capra 290 et al., 2013).

291
In order to better constrain the influence of the channel morphology on the observed erosion/deposition rates, 292 simulations with FLO-2D code were performed (O'Brien et al., 1993). The program routes floods over natural 293 channels solving the full-dynamic wave equation. It has a utility for sediment-transport that can compute 294 sediment scour and deposition. Here the sediment transport capacity equation of Zeller-Fullerton was used 295 (Zeller and Fullerton, 1983;Yang, 1996), (e.g. Fig. 7). For the inflow, the hydrograph of a MPE-type similar   Nat. Hazards Earth Syst. Sci. Discuss., doi:10.5194/nhess-2016-138, 2016 Manuscript under review for journal Nat. Hazards Earth Syst. Sci. Published: 12 May 2016 c Author(s) 2016. CC-BY 3.0 License. pretend to be a qualitative comparison, since absolute depositional rates cannot be compared with those 301 observed on the field. In fact, lahars observed at Volcán de Colima have sediment concentration between 30% 302 and 50%, from hyperconcentrated to debris flow (Vázquez et al., 2014) and depositional rates within these types 303 of flow are higher respect to a flow free of sediment-load (Costa, 1987). The same happens for the erosion, 304 since clear water can be more efficient in eroding the river channel, but lahars commonly induce the collapse 305 of lateral embankments or the entrainment of large blocks from the flow front (Fagents and Baloga, 2006).

306
Despite this assumption, values obtained from the simulation are quite in agreement with those observed in the 307 field (Fig. 7f). In particular, section MG_01 and MG_02 are dominated by erosion (Figs. 7a, and b), mostly 308 deepening the channel, and from MG_03 to MG_05 depositional processes dominates (Figs. 7c, d, and e). The 309 simulation outcome of the 11 June 2013 lahar still presents discrepancy with the results of the field survey but, 310 since as previously stated, it was an extraordinary event. Based on these results, the slope represents the first 311 major factor in controlling E-D rates, as changing from 8° to 5°, depositional processes seems to dominate over 312 erosion (Fig. 7c, d, and e). For example, the ravine morphology at sections MG_01 and MG_04 is quite similar, 313 with a ~80 m-long straight channel, bracketed by two narrow bends, but at MG_04 section, with ~5º in slope, 314 depositional processes are dominating. Depositional process at Montegrande ravine dominate where the 315 channel-bed slope is smaller than 5°, which represents a higher value than that observed in Ruapehu of ~2.7-316 0.7° (Fagents and Baloga, 2006) or the <1.2° found by Pierson (1995) for snow-clad volcanoes; but similar to 317 the values observed in Popocatépetl volcano (i.e. <6.5°, Capra et al., 2004;Muñoz-Salinas et al., 2008, 2009

321
The second parameter that affects erosion and depositional process at Montegrande ravine is the cross section

366
These results can be used as a tool to foresee the effects of future laharic events in Volcán de Colima, and as a 367 tool to improve the input parameters for the modelling of these flows along other volcanoes (e.g., the 497 bedload transport, Geomorphol., 243, 92-105, 2015. 498 Thouret J.C., Oehler, J.F., Gupta, A., Solikhin, A., and Procter, J.N

585
Finally, during the phase of lowest eruptive activity (from 2012-2015), the rate of material available to form lahars 586 was diminishing, while a major quantity of rainfall was needed to trigger the flows.