Effects of the interannual variability of water column stratification on phytoplankton production and biomass in the northern zone off Baja California

A BSTRACT . The ocean off the Baja California Peninsula (Mexico) has been affected by interannual processes. Some of these processes have produced changes in oceanic circulation and the atmosphere, which have been reflected in the structure of the water column. Stratification, defined as the amount of energy needed to mix water throughout the water column, determines nutrient availability in the euphotic zone for phytoplankton growth. The aim of this study is to describe and relate the interannual variations of phytoplankton production and biomass with changes in the structure of the water column. To this end, we analyzed time series (1997–2016) for integrated Chlorophyll a , integrated primary production, pycnocline depth, mixed layer depth, and stratification indices along line 100 of the Investigaciones Mexicanas de la Corriente de California (IMECOCAL) program. The results showed 2 periods of high stratification and a decrease in phytoplankton production and biomass off Baja California, corresponding to the intrusion of subarctic water and El Niño 2015–2016. Finally, by using generalized additive models, we related 2 photosynthethic parameters-maximum photosynthetic rate and maximum light utilization coefficient-with water stratification. These relationships explained interannual variations in phytoplankton production in terms of water column stratification. The time series and the analysis reported here for IMECOCAL line 100 could be used to extrapolate the effects of interannual physical processes on phytoplankton in other zones off the Baja California peninsula.


INTRODUCTION
The ocean is a highly dynamic environment affected by a number of processes that take place at different spatial and temporal scales.The organisms that inhabit there are influenced by local processes, and these processes thus determine, to a great extent, the ecosystem's productivity at a regional level (Daly and Smith 1993).The ocean surface layer is home to marine phytoplankton, the primary component of the food chain.Being autotrophic organisms, their largest abundance occurs in the well-illuminated ocean layer known as the euphotic zone.Given the importance of light in photosynthesis, irradiance is one of the main factors that modulate phytoplankton abundance (Strickland 1965).However, phytoplankton growth and biomass are limited by nutrient availability in the euphotic zone.In this regard, processes at different temporal and spatial scales become important because they determine the structure of the water column, which is directly related to productivity.
One of the major factors limiting the distribution of phytoplankton is water column stratification.Stratification is defined as the amount of energy required to mix water throughout the water column (Simpson and Bowers 1981); if the latter is strongly stratified, more energy will be required to mix it relative to a less stratified column.Stratification of the water column determines the availability of nutrients that are used for phytoplankton growth.The mixed layer depth (MLD) and mixing intensity are physical parameters related to stratification that strongly affect phytoplankton primary production by determining the exposure of phytoplankton to light and its losses by sedimentation (Jäger et al. 2008).A deep MLD leads to a decrease in phytoplankton abundance because the population is distributed throughout a larger volume of water, that is, dilution (Behrenfeld and Boss 2014).In addition, if light conditions change as a result of vertical movements associated with turbulence (deep MLD), phytoplankton will photosynthesize in accordance to the amount of incident light in the MLD (Marra 1978).
The ocean region off the coast of the Baja California peninsula has been affected by interannual processes.Some of these processes have led to changes in ocean circulation, which in turn have modified the structure of the water column.For example, El Niño 1997-1998 produced a deepening of the pycnocline as a result of the influence of tropical warm water transported by a coastal flow parallel to the coast with a poleward direction (Durazo and Baumgartner 2002).During La Niña 1998-1999, intensification of coastal upwelling events led to increased phytoplankton biomass (Lavaniegos et al. 2002).From 2002 to 2006, the weakening of the Alaska Gyre caused the anomalous intrusion of subarctic water off Baja California (Durazo 2009), leading to a decrease in phytoplankton production and biomass (Gaxiola-Castro et al. 2008, Espinosa-Carreón et al. 2015).The ocean warming caused by "The warm Blob" during 2013 and 2014 and El Niño 2015-2016 produced the highest reduction of
El océano frente a las costas de la península de Baja California se ha visto afectado por procesos interanuales.Algunos de estos procesos han generado cambios en la circulación oceánica, lo cual ha modificado la estructura de la columna de agua.Por ejemplo, el evento de El Niño 1997-1998 produjo una profundización en la picnoclina debido la influencia de agua cálida de origen tropical transportada por un flujo costero paralelo a la costa en dirección al polo (Durazo y Baumgartner 2002).Durante La Niña 1998-1999, la intensificación de las surgencias costeras permitió el incremento en la biomasa del fitoplancton phytoplankton biomass and production in the last 13 years at the southern boundary of the California Current (Gómez-Ocampo et al. 2017).However, although phytoplankton production and biomass in this region have responded to these interannual processes, the mechanisms through which these events cause variations in the structure of the water column remain unknown.
Since 1997, the Investigaciones Mexicanas de la Corriente de California (Mexican Investigations on the California Current; IMECOCAL for its acronym in Spanish) program has surveyed the region off Baja California.This has allowed for better knowledge and understanding of the many oceanic physical, biological, and chemical processes in this region (Durazo and Baumgartner 2002, Lavaniegos et al. 2002, 2015;Espinosa-Carreon et al. 2004, Gaxiola-Castro et al. 2008, Gaxiola-Castro 2010, Jeronimo and Gomez-Valdes 2010, Martínez-Gaxiola et al. 2010, Durazo 2015, Espinosa-Carreón et al. 2015, Martinez-Fuentes et al. 2016, Gómez-Ocampo et al. 2017, Lavaniegos 2017).However, although there is available information on physical, biological, and chemical variables, there is still insufficient understanding on how physical processes influence the interannual changes in phytoplankton biomass and production in the water column.
The aim of this article is to describe and relate the interannual variations of phytoplankton biomass and production with changes in the structure of the water column.To this end, the 1997-2016 time series for chlorophyll a, primary production, pycnocline depth, and MLD were analyzed, as well as the stratification indices between the surface and 100 and 200 m ( 100 and  200 , respectively) along IMECOCAL line 100 (L100 IMEC ).L100 IMEC is one of the most intensively sampled lines and, due to its location, it has been found to be representative of the response of the marine ecosystem to interannual processes that occur off the peninsula (Linacre et al. 2010).In addition, in order to understand the relationship between interannual variations in the structure of the water column and phytoplankton production, the relationship between photosynthetic parameters (maximum photosynthetic rate at light saturation [ ] and the maximum light utilization coefficient [ B ]), derived from photosynthesisirradiance experiments conducted at IMECOCAL sampling stations, is explored using stratification indices through generalized additive models (GAMs).

P m B
sensor were downloaded from the ERDDAP database (http://coastwatch.pfeg.noaa.gov/erddap/index.html) for the 2003-2015 period.The primary production (PP) data obtained with the Behrenfeld and Falkowski (1997) Vertically Generalized Production Model (VGPM) was obtained from Oregon State University (OSU, http://www.science.oregonstate.edu/ocean.productivity/).The climatology for the mixed layer MLD, with a spatial resolution of 0.5º × 0.5º, was obtained from the Monthly Isopycnal and Mixed-layer Ocean Climatology (MIMOC, http://www.pmel.noaa.gov/mimoc/) produced by the National Oceanic and Atmospheric Administration (NOAA).Wind data considered for this work were obtained from the National Centers for Environmental Prediction (NCEP).The zonal and southern components (0.25º spatial resolution and 1-day temporal resolution) are available on NOAA's website at https://www.ncdc.noaa.gov.Wind stress curl was calculated from its components according to Trenberth et al. (1990).For all variables, long-period averages (hereafter average conditions) were calculated according to the length of the time period for which data was obtained.From the PP data obtained from OSU, average conditions were calculated for the northern zone off Baja California (28ºN-32ºN) from 1998 to 2002 for data derived from the SeaWiFS sensor and from 2003 to 2016 for data obtained from the MODIS-Aqua sensor (both with a monthly resolution of 18 × 18 km).This climatology was subsequently used to estimate the anomalies from 1998 to 2016, which were defined as the observed value minus the calculated average.
The interannual variations in the structure of the water column and phytoplankton biomass in oceanic and coastal areas were analyzed using data from the stations located along L100 IMEC (Fig. 1).This line was selected because it was the IMECOCAL line representative of the northern zone off Baja California with the largest number of temperature, salinity, and chlorophyll a (Chla) records from 1998 to 2016.Cruise averages were calculated for the variables analyzed at the L100 IMEC stations representative of the coastal zone (stations 30 and 35) and the oceanic zone (stations 45, 50, 40, 55, and60) (Gaxiola-Castro et al. 2010, Durazo 2015).Phytoplankton Chla samples were obtained from water samples collected with Niskin bottles at standard depths (0, 10, 20, 50, and 100 m) and analyzed with the fluorimetric method (Yentsch andMenzel 1963, Holm-Hansen et al. 1965).Temperature and salinity CTD (Sea-Bird) data were used to estimate MLD following the method proposed by Kara et al. (2000) and modified by Jeronimo and Gomez-Valdes (2010) for the IMECOCAL zone.Pycnocline depth (Z Pyc ) was calculated according to Fiedler et al. (2013).The stratification parameter () between the surface and 100 m and the surface and 200 m depth ( 100 and  200 , respectively) was calculated according to Simpson and Bowers (1981).Chla was integrated into the water column from the surface to 100 m depth using the trapezoidal rule to estimate integrated Chla (Chla int , mg•m -2 ).
Los experimentos fotosíntesis-irradiancia (curvas P-E) se realizaron a bordo en las estaciones oceánicas de algunas campañas oceanográficas, con muestras de agua recolectadas entre las 10:00 y las 14:00, tiempo local, en la profundidad correspondiente al 30% de irradiancia superficial.El agua fue inside of each bottle in the incubator with a QSL 100 (Biospherical Instruments).In order to represent the interannual variation in photosynthetic parameters in the northern zone off Baja California and given the scarcity of the P-I experiments for L100 IMEC ,  B and averages were calculated for the stations of IMECOCAL lines 100, 103, 110, and 113.The criterion for using this average stems from the fact that the variability in the physical and biological conditions in the region north of Vizcaíno Bay (28ºN) is spatially homogeneous in response to the different temporal scales (Gaxiola-Castro et al. 2010, Durazo 2015).Therefore, phytoplankton communities in this area would be expected to be similar.The information on the methods and equipment used during the IMECOCAL cruises is detailed in http://imecocal.cicese.mx/.

Statistical analysis
The relationship of trends PP and water column structure was explored using GAMs (Hastie and Tibshirani 1986).This method fits non-linear relationships between variables, which are expressed as a curve from a smoothed function.In contrast with linear models, GAMs allow assigning an exponential distribution (Poisson, binomial, gamma, or normal) to the dependent variable.
In order to explain the variation in PP in terms of phytoplankton physiology, the  B and parameters were related to structure of the water column represented by  100 .To obtain a more robust statistical fit, P-I experiments carried out between 1998 and 2013 were used (no experiments were conducted in 2014 and 2016) at different stations of the IME-COCAL lines (n = 273).The relationships were fitted with the R program for statistical calculations using the ''mixed GAM computation vehicle'' (mgcv) package (Wood et al. 2015).The mgcv package sets curve functions to the terms of the model and makes a cross-validation to determine the optimal smoothing degree (i.e., degrees of freedom) (Wood 2006).The best distribution was assigned to the predictor variables  B and using the -square goodness-of-fit test.

Average conditions in the study area
In order to characterize the horizontal distribution of some variables related to water column structure and productivity off Baja California, average conditions in the study area were obtained for MLD, SST, wind stress curl, and surface chlorophyll (Fig. 1).Average MLD and SST show higher values offshore and in the oceanic area to the south of the Baja California peninsula (Fig. 1a, b).In contrast, average wind stress curl and chlorophyll values derived from satellite imagery were higher in the coastal zone and decreased offshore and to the south of the peninsula (Fig. 1c, d).Therefore, 2 typical average conditions were observed in the study area.recolectada con botellas Niskin de 5 L de capacidad para llenar 27 botellas planas de poliestireno de 250 mL (Nucleon), las cuales se inocularon con 100 L de NaH 14 CO 3 (~5 Ci).Las botellas fueron colocadas en un incubador tipo Morel (Marcel et al. 1994) por ~2 h en un gradiente de luz de 1 a 900 mol•cuanta•m -2 •s -1 generado por una lámpara de tungsteno-halógeno de 500 W. La tasa de asimilación de carbono (P, mgC•m -3 •h -1 ) se obtuvo a partir de la incorporación de 14 C durante la incubación, sustrayendo los valores de tiempo cero, y se normalizó con respecto a la Chla medida a la profundidad del 30% de irradiancia superficial en cada experimento ) fueron calculados con la función hiperbólica descrita por Jassby y Platt (1976): , donde E L es la irradiancia de la lámpara de tungstenohalógeno medida en el interior de cada botella en el incubador, con un QSL 100 de Biospherical Instruments.Con el fin de representar la variación interanual en los parámetros fotosintéticos en la zona norte frente a Baja California y debido a la escasez de experimentos P-E en la L100 IMEC , se calcularon los promedios de  B y , para las estaciones de las líneas 100, 103, 110 y 113 del IMECOCAL.El criterio para realizar este promedio se basó en que se ha reportado que la variabilidad de las condiciones físicas y biológicas en la región al norte de bahía Vizcaíno (28ºN) son espacialmente homogéneas en respuesta a las diferentes escalas temporales (Gaxiola-Castro et al. 2010, Durazo 2015).Por lo tanto, se esperaría que las comunidades de fitoplancton presentes en esta zona sean similares.La información de los métodos y equipos usados en los cruceros del programa IMECOCAL se encuentra detallada en http://imecocal.cicese.mx/.
Con el fin de explicar la variación en la PP en términos de la fisiología del fitoplancton, los parámetros  B y fueron relacionados con la estructura de la columna de agua representada por  100 .Para obtener un ajuste estadístico más robusto, se usaron experimentos P-E llevados a cabo entre 1998-2013 (no hubo experimentos en 2014 y 2016) en diferentes estaciones de las líneas del programa IMECOCAL (n = 273).Las relaciones fueron ajustadas con el programa R para cálculos estadísticos usando el paquete "mixed GAM computation vehicle" (mgcv) (Wood et al. 2015).El paquete mgcv ajusta funciones de curvas a los términos del modelo y The first occurs in the coastal area and off the northern part of the peninsula, and seems to be more dynamic according to the wind stress curl values; consequently, the water column is expected to be less stratified.The second, which is less dynamic, occurs mainly in the oceanic area and off the southern part of the peninsula, with conditions typical of a more stratified environment.

1997-2016 time series in L100 IMEC
The effect of the interannual variability in water column structure on phytoplankton was characterized from time series of the variables of interest in the L100 IMEC coastal and oceanic areas.Particular periods during which the Chla int signal showed obvious variations were selected (Fig. 2a).These periods correspond to the period from 2002 to 2006 (anomalous intrusion of subarctic water), La Niña 2008, and the ocean warming during 2013 to 2016 resulting from "The warm Blob" and El Niño 2015-2016.
The time series of the annual Chla int average showed a clear response to interannual events in coastal and oceanic areas (Fig. 2a).The lowest Chla int values (< 40 mg m -2 ) in the coastal and oceanic areas occurred between 2002 and 2007.In contrast, La Niña 2008 was associated with the maximum Chla int values of the time series in both areas, with similar values in the coastal (~76 mg•m -2 ) and oceanic (~75 mg•m -2 ) areas.From 2008 to late 2016 there was a decline in Chla int values in the oceanic area, reaching minimum values as low as ~30 mg•m -2 but slightly higher compared to those observed during the subarctic water intrusion (2002 to 2006).In 2013 Chla int remained relatively unchanged until 2015, when it increased.In the coastal area the lowest Chla int values were recorded between 2011 and 2014 and were similar to those observed during the subarctic water intrusion.In this area, Chla int increased in 2014 until reaching ~45 mg•m -2 in 2016.
Phytoplankton production in the northern zone off Baja California showed periods with increasing and decreasing values through the time series (Fig. 2b).During the intrusion of subarctic water, no definite trend was observed.However, conditions during La Niña 2008 led to a rise in PP rates.Afterwards, due to the "The warm Blob" effect during 2014, phytoplankton production dropped and attained even lower levels during El Niño 2015-2016.
MLD was deeper in the oceanic than in the coastal area, but variation was similar in both areas from 1998 to 2016 (Fig. 2c).Aspects worth noting are the reduction in the thickness of the mixing layer during the subarctic water intrusion and the particularly noticeable deepening in the oceanic area in 2004.A period of shallow MLD also took place during La Niña 2008, reaching a minimum value in 2011 in the coastal and oceanic areas (12 and 25 m, respectively).However, the most superficial MLD values in both areas occurred from 2010 to 2012, after which the mixed layer became deeper in hace validación-cruzada para determinar el grado óptimo de suavizado (i.e., grados de libertad) (Wood 2006).Se asignó la mejor distribución a las variables predictoras  B y usando la prueba de bondad de ajuste  cuadrada.

RESULTADOS Condiciones promedio del área de estudio
Con el fin de caracterizar la distribución horizontal de algunas variables relacionadas con la estructura de la columna de agua y la productividad frente a Baja California, se obtuvieron las condiciones promedio en el área de estudio para la MLD, TSM, rotacional del esfuerzo del viento y clorofila superficial (Fig. 1).Las condiciones promedio de la MLD y la TSM presentan valores más altos hacia afuera de la zona costera y en la zona oceánica hacia el sur de la península de Baja California (Fig. 1a, b).Contrario a lo anterior, los valores promedio del rotacional del esfuerzo del viento y de clorofila derivada de satélite son más altos en la zona costera y disminuyen hacia mar abierto y hacia el sur de la península (Fig. 1c, d).Por lo tanto, en el área de estudio se observó que ocurren 2 condiciones promedio típicas.La primera ocurre en la zona costera y frente al norte de la península, que según los valores del rotacional del esfuerzo de viento parece ser más dinámica y, por ende, se espera que la columna de agua sea menos estratificada.La segunda, menos dinámica, ocurre principalmente en la zona oceánica y frente al sur de la península e implica condiciones típicas de un ambiente más estratificado.
Pycnocline depth (Z Pyc ) was more variable in the oceanic than in the coastal area (Fig. 2d).In 2004, during the subarctic water intrusion, there was a slight reduction in Z pyc in the oceanic area (~40 m).The same behavior was observed during 2010, when Z Pyc reached its lowest in both areas (~30 and 45 m in the coastal and oceanic areas, respectively).However, the deepest Z Pyc value (~90 m) was recorded in the oceanic area in 2012, prior to "The warm Blob".From that year, the pycnocline in this zone became shallower and reached the value closest to the surface in 2014 (~30 m).Subsequently, with the onset of El Niño, Z Pyc deepened in 2015 and reached its maximum value in 2016, with a value similar to the one observed in 2007 (~80 m, first peak in the mente más altos respecto a aquellos observados durante la intrusión de agua subártica (2002 a 2006).En 2013 no se observó variación significativa en la Chla int hasta 2015, cuando incrementó.En la zona costera, los valores más bajos de Chla int ocurrieron entre 2011 y 2014 y fueron similares a los observados durante la intrusión del agua subártica.En esta zona, a partir de 2014, la Chla int se incrementó hasta llegar a un valor de ~45 mg•m -2 en 2016.
La producción del fitoplancton en la zona norte frente a Baja California mostró periodos de aumento y disminución a lo largo de la serie de tiempo (Fig. 2b).Durante la intrusión de agua subártica, no se observó una tendencia definida.Sin embargo, las condiciones durante La Niña 2008 generaron el aumento en las tasas de PP.Enseguida, debido al efecto de "The Warm Blob" durante 2014, se redujo la producción del  -Aqua (2003-Aqua ( -2016)), (c) profundidad de la capa de mezcla ( MLD, m), (d) profundidad de la picnoclina (Z Pyc , m) y parámetro de estratificación de la columna de agua a (e) 100 y (f) 200 m de profundidad ( 100 and  200 , respectivamente; Joules) a lo largo de la línea 100 del IMECOCAL.Los colores indican la serie de tiempo para estaciones costeras (rojo) y oceánicas (azul).Las líneas punteadas de color denotan las medias mensuales y las líneas de color sólidas indican las medias anuales.Las líneas grises punteadas indican los periodos seleccionados para el análisis.time series).Although variations were not as marked in the coastal area as in the oceanic area, an aspect worth noting is the slight deepening of the pycnocline during the subarctic water intrusion in 2002 and from 2012 to 2016, under the influence of "The warm Blob" and El Niño.
The stratification indices  100 and  200 showed a similar variation in both areas (Fig. 2e, f).The time series for both indices reveal an increase in the stratification of the water column during the intrusion of subarctic water (up to  100 150 J•m -3 in both areas, and  200 ~ 450 J•m -3 and  200 350 J•m -3 in the oceanic and coastal areas, respectively) and the peak observed in 2015 after the "The warm Blob" and the onset of El Niño 2015-2016 ( 100 ~ 225 J•m -3 and  200 > 500 J•m -3 in both areas).

DISCUSSION
Of all the variables analyzed, phytoplankton biomass was the one that best projected the interannual variability due to the various events that influenced the ocean off the northern part of Baja California.Our hypothesis is that highstratification periods can reduce phytoplankton biomass and affect phytoplankton photophysiology (decreased  B , increased ), and these changes are in turn reflected in the variations in PP.
Stratification was associated with decreasing and increasing phytoplankton production and biomass during the intrusion of subarctic water, La Niña 2008, and the ocean warming caused by "The warm Blob" and El Niño 2015-2016.A marked water column stratification was observed during the warm periods and the intrusion of subarctic water, coupled with the reduction in MLD and Z Pyc .In a poorly mixed water column, phytoplankton can remain in the upper, well-illuminated layer during sufficient time to maintain a population (Huisman 1999).In contrast, when the water column is well mixed, phytoplankton is transported throughout it and, in time, each microorganism experiences the depth-averaged light intensity, which decreases according to MLD.As a result, the depth-averaged specific PP drops as the depth of the water column increases (Huisman 1999, Diehl et al. 2002).In addition, MLD and mixing intensity are related to the accumulation of biomass through the losses of phytoplanktonic cells by sedimentation (Jäger et al. 2008).The probability that a phytoplanktonic organism or a colony sinks below the euphotic zone decreases as mixing intensity increases (Huisman et al. 2004).This occurs because turbulent mixing disperses phytoplankton throughout the water column, which partially offsets sedimentation.In general, losses by sinking affect primarily organisms that sink rapidly (e.g., diatoms) in shallow and poorly mixed water columns (Diehl et al. 2002).In summary, a poorly mixed water column (high stratification) favors PP rates but restrains the accumulation of biomass, as observed in the period of subarctic water intrusion between 2003 and 2005 and during "The warm Blob" in 2013-2014 (Fig. 2).La MLD fue más profunda en la zona oceánica, pero presentó una variación similar en ambas zonas desde 1998 hasta 2016 (Fig. 2c).Destacan la reducción del espesor de la capa de mezcla durante la intrusión de agua subártica y la profundización particularmente notoria en 2004 en la zona oceánica.Durante el evento de La Niña 2008 también hubo un periodo de disminución en la MLD, la cual llegó a un valor mínimo en 2011 en las zonas costera y oceánica (12 y 25 m, respectivamente).Sin embargo, los valores de MLD más superficiales en ambas zonas ocurrieron desde 2010 hasta 2012; a partir de este último año la capa de mezcla se profundizó en la zona oceánica hasta alcanzar un máximo de ~50 m en 2013 bajo la influencia de "The Warm Blob".

DISCUSIÓN
De las variables analizadas, la biomasa de fitoplancton fue la que mejor reprodujo la variabilidad interanual debida a diferentes eventos que influenciaron el océano frente a la zona norte de Baja California.Nuestra hipótesis es que los periodos de alta estratificación causan la reducción en la biomasa del fitoplancton y afectan la foto-fisiología del The largest reduction in PP rates was observed during El Niño 2015-2016, despite the fact that stratification in the northern region off Baja California was the highest in the past 19 years.This may have occurred because, for the first time in history, there was a warm event that preceded an El Niño event (Jacox et al. 2016).The marked deepening of the pycnocline and of the mixed layer during "The warm Blob", coupled with the high water column stratification during El Niño 2015 (Figs.2d-f), probably limited the availability of nutrients in the euphotic zone, leading to a decrease in phytoplankton growth rates and hence resulted in lower PP.In addition, the warming of the upper ocean layer (0-100 m) prompted by both events limited phytoplankton photosynthesis.
The depth of the pycnocline and the mixed layer showed greater variability in the oceanic area.This may be due to the prevalence of remote forcings over local forcings (Durazo et al. 2017).In general, in addition to remote signals, the coastal environment is also influenced by local forcings that can mitigate temporal variability.Therefore, Z Pyc and MLD in the oceanic area best represent the interannual variability in the northern zone off Baja California.
In order to assess the relationship between the variability of phytoplankton production and stratification, GAMs were used to determine the relationship between 2 of the characteristic phytoplankton photosynthetic parameters,  B and , and the  100 stratification index (Fig. 3).
is associated with enzymatic photosynthetic processes and depends on factors such as temperature (Eppley 1972), nutrient concentration (Glover 1980) and cell size (Malone 1980), among others.The  B parameter depends on cell size (Taguchi 1981), pigment composition (Handall 1970), and nutrient availability (Platt et al. 1992).In this work,  B showed high values when the water column was more mixed (low  100 values, Fig. 3a), whereas increased along with stratification (increase in  100 , Fig. 3b).These relations explained the interannual variations in PP in terms of phytoplankton physiology in the northern region off Baja California.
The above results suggest that PP rates are influenced by stratified environments (Huisman 1999, Diehl et al. 2002).This relationship is supported by the analysis with GAMs, revealing that high stratification is related to peak maximum photosynthetic rates.In addition, it is supported by the temporal variation in the parameter, which peaked in 2003, when increased stratification and PP co-occurred (Fig. 4a).The increased could be related to the photoacclimation of small phytoplankton cells-such as cyanophyta and prochlorophyta-at high irradiance levels, as suggested by the results of Geider et al. (1993); these authros observed changes in between phytoplankton of different cell sizes, with low in large diatoms at the same growth irradiance.On the other hand,  B was higher in turbulent environments (well-mixed water column, less stratification), likely due to the presence of large phytoplankton cells that were photoacclimated at low ), y estos cambios inciden en las variaciones de la PP.
La profundidad de la picnoclina y la capa de mezcla presentaron mayor variabilidad en la zona oceánica.Esto puede deberse a la importancia relativa de los forzamientos remotos P m B irradiances, since they are displaced throughout the water column.These differences in magnitude due to phytoplankton size had been previously reported for the same study area by Gonzalez-Morales et al. (1993).Sosa-Avalos et al. (forthcoming 2017 Sep) found that the variability in  B resulted from spatial and temporal changes in phytoplankton populations, which were dominated mostly by diatoms and dinoflagellates during spring.It is precisely at this time of the year when the water column is poorly stratified due to the presence of more intense winds and coastal upwelling events (Durazo 2015).This was observed in 2007, the year that showed poor stratification coupled with maximum PP rates, an increase in biomass, and the highest  B value in the time series (Figs. 2 y 4b).Therefore, high  B in periods of reduced stratification suggests the presence of large phytoplankton species, while high in periods of marked stratification could be related to the presence of smaller phytoplankton species.
The results of this study support the conclusion that the periods of strongest water column stratification were the subarctic water intrusion from 2002 to 2006 and El Niño 2015-2016, which resulted in the greatest decline in phytoplankton biomass (in both periods) and PP (during El Niño 2015-2016) in the northern zone off Baja California.The photosynthetic parameters  B and associated with phytoplankton photophysiology explained the variations in PP in relation to the stratification of the water column.The time series and the analysis presented here for L100 IMEC will serve as indicators of the physical processes that occur off the peninsula and their implications for phytoplankton.These  (Durazo et al. 2017).De manera general, el ambiente costero recibe, además de señales remotas, la influencia de forzamientos de escala local que pueden reducir la variabilidad temporal.Por lo tanto, la Z Pyc y la MLD en la zona oceánica representan mejor la variabilidad interanual en la zona norte frente a Baja California.
A través de este estudio se concluye que los periodos de mayor estratificación en la columna de agua fueron la   California.Los parámetros fotosintéticos  B y relacionados con la foto-fisiología del fitoplancton explicaron las variaciones en la PP en relación con la estratificación de la columna de agua.Las series de tiempo y el análisis presentado aquí para la L100 IMEC servirán como indicadores de los procesos físicos que ocurren frente a la península y sus consecuencias para el fitoplancton, los cuales pueden ser extrapolados hacia otras zonas del IMECOCAL para explicar la variabilidad anual e interanual de la biomasa y PP del fitoplancton.
de  B , incremento de

Figure 3 .
Figure 3. Partial response of (a) the maximum light utilization coefficient( B , mg C•[mg Chla] -1 •h -1 •[mol•quanta•m -2 •s -1 ] -1 )) and (b) maximum photosynthetic rate ( , mg C•[m Chla•h] -1) to the water column stratification parameter at 100 m depth ( 100 , a and b), as defined by the generalized additive models (GAMs).The smoothing functions (S) are represented as solid lines and the 95% confidence Figure 3. Partial response of (a) the maximum light utilization coefficient( B , mg C•[mg Chla] -1 •h -1 •[mol•quanta•m -2 •s -1 ] -1 )) and (b) maximum photosynthetic rate ( , mg C•[m Chla•h] -1 ) to the water column stratification parameter at 100 m depth ( 100 , a and b), as defined by the generalized additive models (GAMs).The smoothing functions (S) are represented as solid lines and the 95% confidence intervals as dashed lines.Rugged lines on the x-axes represent the observed  100 values.The y-axis labels show the GAM smoothing function.The vertical dashed lines represent the threshold values of the change from positive to negative (or viceversa) influence of the predictor variable ( 100 ) on the response variable ( B ,).
extrapolated to other IMECOCAL areas to explain the annual and interannual variability of phytoplankton biomass and PP.

Figure 4 .
Figure 4. Time series for the average (a) maximum photosynthetic rates ( , mg C•[m Chla•h] -1 ) and (b) the maximum light utilization coefficient( B , mg C•[mg Chla] -1 •h -1 •[mol•quanta•m -2 •s -1 ] -1 )at the oceanic stations of the IMECOCAL program.Dashed lines indicate the peak and  B values of the time series.