Commit 1d003d51 authored by DORAY's avatar DORAY
Browse files
parents 8be2e9ef e32df760
---
title: "Space-time distribution of clupeiforms in the in the Gulf of Lions in the summertime"
subtitle: "DEFIPEL project"
author: Tarek Hattab, Ifremer / MARBEC
date: "08/09/2021"
abstract: "This document summarises the results on the space-time distribution and habitat of clupeiforms in the Gulf of Lions in the summertime from 2003 to 2019 as observed during the PELMED survey."
header-includes:
- \usepackage{float}
- \floatplacement{figure}{H}
output:
bookdown::pdf_document2: default
bookdown::html_document2: default
toc: yes
toc-depth: 10
# pdf_document:
# extra_dependencies: flafter
# html_document:
# df_print: paged
---
```{r setup, include=FALSE}
knitr::opts_chunk$set(echo = FALSE)
#,fig.pos = "!H", out.extra = "")
library(knitr)
library(bookdown)
pref='G:/Tarek/Ifremer/Papiers/DEFIPEL_MFA_V2/Fish/MFA_Results/'
path.export.rasters.SPF=pref
```
\newpage
# Introduction
- Objectives:
- analyse the summertime distribution of clupeiforms in Gulf of Lions in space and time, based on survey (PELMED) data
- describe main spatial patterns and temporal trends in clupeiforms distribution in the summertime from 2003 to 2019
- relate clupeiforms communities to concurrent hydrological conditions to define habitats.
# Material and methods
- Data
- 0.1° gridmaps of 2003-2019 clupeiforms biomass (anchovy, sardine, sprat), per length class from PELMED survey. The biomasses were aggregated by considering 4 evenly spaced length class intervals for each species as follows:
Anchovy: [2,6] ,(6,10], (10,14] and (14,18.5] cm
Sardine: [5,9], (9,13], (13,17] and (17,21] cm
Sprat: [5,7],(7,10], (10,12], and (12,14.5] cm <br />
- 0.1° gridmaps of 2003-2019 hydrological variables were derived from PELMED survey CTD casts. Seven hydrological indices were computed at each CTD station to assess the hydrological conditions experienced by small pelagic fish: surface and bottom temperature (SurfTemp and BotTemp), surface salinity (SurSal), a river plume index, the equivalent freshwater height (Fwh), and water column stratification indices: the potential energy deficit (Dpe), the mixed layer maximum depth (mixZ), and the pycnocline depth (pycZ). Two variables derived from satellite observations were added including a July climatology of Sea Surface Temperature (SST_SAT) and chlorophyll-a concentration (from OC5 bio-optical algorithm) (Chla).
- Space-time MFA (\~ grouped PCA) on series of clupeiforms biomass gridmaps (FactoMineR package)
- Rows: gridmaps cells
- Columns: Clupeiforms biomass per length class
- Groups: years
- Space-time MFA (~ grouped PCA) on series of hydrological variables
– Rows: gridmaps cells
– Columns: hydrological variables
– Groups: years
- MFA results interpreted in:
- MFA space: individuals (cells) / variables / groups (time)
- Geographical space: MFA maps
- Definition of habitats of clupeiforms communities:
- To investigate the relationships between the main gradients identified in the MFA ordination spaces for fish and hydrology, map cells coordinates on selected fish MFA components were projected as supplementary variables in the hydrology MFA compromise space. The correlations between fish and hydrology selected MFA components in the hydrology ordination space were analyzed to assess which hydrological factors may influence the spatial distribution of the small pelagic fish communities.
# Results
## Clupeiforms size distribution
### Deviance explained by MFA axis
```{r fig1, echo=FALSE, out.width="50%",fig.cap='Deviance explained'}
#path.export.results.clup=paste(pref,'PELGAS/Resultats/gridMaps/MFA/Poissons/2000-2019/Clupeiforms/',sep='')
#load(file=paste(path.export.results.clup,"MFAclupResults.RData",sep=''))
print(paste(174,'map cells x',165,'variables',
'grouped into',17,'years'))
#NMFA95=dimnames(res.fssd.MFA.clup$eig[round(res.fssd.MFA.clup$eig[,3])>=95,])[[1]]
print(paste('95% deviance explained by',38,'MFA components'))
#barplot(res.fssd.MFA.clup$eig[,3],main="Cumulative % of variance explained",
# names.arg=1:nrow(res.fssd.MFA.clup$eig))
knitr::include_graphics(paste(path.export.rasters.SPF,'Cumulative_Variance explained.png',sep=''))
```
### MFA individuals (cells) plane
Map cells in MFA1:3 planes are presented in Figure \@ref(fig:fig2):
```{r fig2, echo=FALSE, out.width="100%",message=FALSE,fig.cap="Map cells in MFA1:3 plane"}
#library(factoextra)
#library(gridExtra)
#p1 = fviz_mfa_ind(res.fssd.MFA.clup, col.ind = "contrib",title='')
#p2 = fviz_mfa_ind(res.fssd.MFA.clup, col.ind = "contrib",axes = c(2, 3),title='')
#grid.arrange(p1, p2, ncol = 2)
knitr::include_graphics(paste(path.export.rasters.SPF,'individuals in MFA1-2-3 plane.png',sep=''))
```
### Variables contributions to MFA axes
```{r fig3, echo=FALSE, out.width="100%",fig.cap = "Variables with abs(correlation) with MFA1 higher or equal to 0.5. Years in top pannel, species as rows, size categories in bottom pannel."}
knitr::include_graphics(paste(path.export.rasters.SPF,'Figure_fishMFA123corVar_V2.png',sep=''))
```
Variables well correlated with MFA1 (Figure \@ref(fig:fig3)):
- MFA1 positive correlation with small/medium sardine ([5,9] and (9,13] cm) and with small/medium sprat ([5,7] and (7,10] cm)
- MFA1 negative correlation with large anchovy of (14,18.5] cm
Variables well correlated with MFA2 (Figure \@ref(fig:fig3)):
- MFA2 positive correlation with large sprat of (12,14.5] from 2003 to 2008, and medium sprat of (10,12] cm from 2008.
- Positive correlation with big sardine of (17,21] cm at the beginning of the time series from 2003 to 2008.
### Clupeiforms mean spatial patterns: MFA maps
```{r fig4, echo=FALSE, out.width="90%",fig.cap = "MFA1, 2 and 3 loadings maps."}
knitr::include_graphics(paste(path.export.rasters.SPF,'rasterLevelPlot_MFAcoordGeoSpace.png',sep=''))
```
Three main spatial gradients in MFA1 2 and 3 loadings (Figure \@ref(fig:fig4)):
- MFA1: Coastal (positive) vs offshore areas (negative) decreasing gradient
- MFA2: central part of the Gulf (positive) vs. the remaining areas (negative)
- MFA3: Western area (positive) VS Eastern area/mouth of the Rhone (negative)
### Clupeiforms spatial patterns inter-annual variability
```{r fig5, echo=FALSE, out.width="90%",fig.cap = "Inter-annual inertia of MFA1 and 2 loadings."}
knitr::include_graphics(paste(path.export.rasters.SPF,'rasterLevelPlot_1-3_MFAcoordInertia.png',sep=''))
```
Inter-annual variability (inertia) seems relatively homogeneous over the entire Gulf (Figure \@ref(fig:fig5)).
### Characteristic areas and definition of clupeiforms communities
Clupeiforms communities were defined by analysing the species composition in the different areas defined along MFA gradients in the previous step.
```{r fig6, echo=FALSE, out.width="90%",message=FALSE,fig.cap="Characteristic areas occupied by clupeiforms communities observed in the summertime in the Gulf of lion"}
knitr::include_graphics(paste(path.export.rasters.SPF,'rasterLevelPlot_MFAcoordMaskCos2GeoSpace.png',sep=''))
```
```{r fig7, echo=FALSE, out.width="90%",message=FALSE,fig.cap="Biomass distribution of clupeiforms communities encountered in the Gulf of Lion characteristic areas (MFA1)."}
knitr::include_graphics(paste(path.export.rasters.SPF,'Figure_fis_clumps_MFA1.png',sep=''))
```
```{r fig8, echo=FALSE, out.width="90%",message=FALSE,fig.cap="Biomass distribution of clupeiforms communities encountered in the Gulf of Lion characteristic areas (MFA2)."}
knitr::include_graphics(paste(path.export.rasters.SPF,'Figure_fis_clumps_MFA2.png',sep=''))
```
```{r fig9, echo=FALSE, out.width="90%",message=FALSE,fig.cap="Biomass distribution of clupeiforms communities encountered in the Gulf of Lion characteristic areas (MFA3)."}
knitr::include_graphics(paste(path.export.rasters.SPF,'Figure_fis_clumps_MFA3.png',sep=''))
```
### Clupeiforms MFA annual anomalies map
```{r fig10, echo=FALSE, out.width="70%",fig.cap = "MFA1 annual anomalies map, 2003-2019."}
knitr::include_graphics(paste(path.export.rasters.SPF,'rasterLevelPlot_MFA1pcoordGeoSpace.png',sep=''))
```
```{r fig11, echo=FALSE, out.width="70%",fig.cap = "MFA2 annual anomalies map, 2003-2019."}
knitr::include_graphics(paste(path.export.rasters.SPF,'rasterLevelPlot_MFA2pcoordGeoSpace.png',sep=''))
```
```{r fig12, echo=FALSE, out.width="70%",fig.cap = "MFA3 annual anomalies map, 2003-2019."}
knitr::include_graphics(paste(path.export.rasters.SPF,'rasterLevelPlot_MFA3pcoordGeoSpace.png',sep=''))
```
## Distribution of hydrological variables
### Deviance explained by MFA axis
```{r fig13, echo=FALSE, out.width="50%",fig.cap='Deviance explained'}
#path.export.results.clup=paste(pref,'PELGAS/Resultats/gridMaps/MFA/Poissons/2000-2019/Clupeiforms/',sep='')
#load(file=paste(path.export.results.clup,"MFAclupResults.RData",sep=''))
print(paste(174,'map cells x',153,'variables',
'grouped into',17,'years'))
#NMFA95=dimnames(res.fssd.MFA.clup$eig[round(res.fssd.MFA.clup$eig[,3])>=95,])[[1]]
print(paste('95% deviance explained by',24,'MFA components'))
#barplot(res.fssd.MFA.clup$eig[,3],main="Cumulative % of variance explained",
# names.arg=1:nrow(res.fssd.MFA.clup$eig))
path.export.rasters.hydro<-"G:/Tarek/Ifremer/Papiers/DEFIPEL_MFA_V2/Hydro/MFA_Results/"
knitr::include_graphics(paste(path.export.rasters.hydro,'Cumulative_Variance explained.png',sep=''))
```
### MFA individuals (cells) plane
Map cells in MFA1:2 planes are presented in Figure \@ref(fig:fig14):
```{r fig14, echo=FALSE, out.width="100%",message=FALSE,fig.cap="Map cells in MFA1:2 plane"}
knitr::include_graphics(paste(path.export.rasters.hydro,'individuals in MFA1-2 plane.png',sep=''))
```
### Variables contributions to MFA axes
```{r fig15, echo=FALSE, out.width="100%",fig.cap = "Variables with abs(correlation) with MFA1 higher or equal to 0.6. Years in top pannel, species as rows, size categories in bottom pannel."}
knitr::include_graphics(paste(path.export.rasters.hydro,'Figure_hydroMFA123corVar_V2.png',sep=''))
```
Variables well correlated with MFA1 (Figure \@ref(fig:fig15)):
- MFA1 positive correlation with DefEpot, Fwh and mixZ
- MFA1 negative correlation with BotTemp and Chla
Variables well correlated with MFA2 (Figure \@ref(fig:fig15)):
- MFA2 positive correlation with SST_SAT and SurfTemp
### Mean spatial patterns of hydrological conditions: MFA maps
```{r fig16, echo=FALSE, out.width="100%",fig.cap = "MFA1, and 2 loadings maps."}
knitr::include_graphics(paste(path.export.rasters.hydro,'rasterLevelPlot_MFAcoordGeoSpace.png',sep=''))
```
Two main spatial gradients in MFA1 & 2 loadings (Figure \@ref(fig:fig16)):
- MFA1: Coastal (negative) vs offshore areas (positive) increasing gradient
- MFA2: Western area (negative) VS Eastern area/mouth of the Rhone (positive)
## Definition of habitats of clupeiforms communities
```{r fig17, echo=FALSE, out.width="90%",fig.cap = " Correlation between components of the Multiple Factor Analysis (MFA1-3) of fish biomass and hydrological variables (MFA1-2)."}
knitr::include_graphics(paste(path.export.rasters.SPF,'Fish_Hydro.png',sep=''))
```
The first variance component of the fish MFA (MFA1 fish) was significantly negatively correlated with the first variance component of the hydrology MFA (MFA1 hydro)
The third variance component of the fish MFA (MFA3 fish) was significantly negatively correlated with the second variance component of the hydrology MFA (MFA2 hydo)
# Conclusions
- Main Clupeiforms spatial pattern (26 % deviance explained): MFA1 coastal vs offshore areas negative gradient.
- More small/medium sardine near the coast since 2002
- Less large sardine (13,17] in coastal area from 2009
- More small/medium sprat in coastal area since 2002
- More large anchovy offshore and small anchovy near the coast
- Secondary Clupeiforms spatial pattern (14.7% deviance explained): MFA2 central part of the Gulf (positive) vs. northern and southern areas (negative).
- More medium size anchovy (10,14] in the central area since 2002
- More large sprat of (12,14.5] from 2003 to 2008, and medium sprat (10,12] from 2008 in the central area
- More big sardine of (17,21] in the central area from 2003 to 2008.
- Thirdly Clupeiforms spatial pattern (7% deviance explained): MFA3 Western area (positive) VS Eastern area (negative).
- More medium and big size anchovy and sardine in western area
- Main hydrology spatial pattern (36.8 % deviance explained): MFA1 Coastal (negative) vs offshore areas (positive) increasing gradient.
- less stratified and more mixed water near the coast
- More stratification offshore
- More chlorophyll-a in coastal area
- Higher Sea bottom temperature in coastal area
- Secondary hydrology spatial pattern (18.7 % deviance explained): MFA2: western area (negative) VS eastern area/mouth of the Rhone (positive).
- Higher Sea surface temperature in eastern area
- The first variance component of the fish MFA was significantly negatively correlated with the first variance component of the hydrology MFA.
- Communities were distributed along the onshore–offshore gradient in productivity and water column stratification
- The second variance component of the fish MFA (MFA2) was not correlated with any of the hydrology MFA components.
- This indicates that the spatial distribution of the MFA2 associated community was not likely to be shaped by the considered hydrological variable
- The third variance component of the fish MFA was significantly negatively correlated with the second variance component of the hydrology MFA.
- Communities were distributed along the west-est Sea surface temperature gradient with bigger sardine and anchovy in coldest water (west)
# Annexes: gridmaps
## Fish gridmaps
```{r, echo=FALSE, out.width="90%",fig.cap = "ENGR-ENC-[2,6]"}
gridmaps.path<-'G:/Tarek/Ifremer/Papiers/DEFIPEL_MFA_V2/Fish/Rasters/Plots/'
knitr::include_graphics(paste(gridmaps.path,"ENGR-ENC-[2,6].png",sep=''))
```
```{r, echo=FALSE, out.width="90%",fig.cap = "ENGR-ENC-(6,10]"}
knitr::include_graphics(paste(gridmaps.path,"ENGR-ENC-(6,10].png",sep=''))
```
```{r, echo=FALSE, out.width="90%",fig.cap = "ENGR-ENC-(10,14]"}
knitr::include_graphics(paste(gridmaps.path,"ENGR-ENC-(10,14].png",sep=''))
```
```{r, echo=FALSE, out.width="90%",fig.cap = "ENGR-ENC-(14,18.5]"}
knitr::include_graphics(paste(gridmaps.path,"ENGR-ENC-(14,18.5].png",sep=''))
```
```{r, echo=FALSE, out.width="90%",fig.cap = "SARD-PIL-[5,9]"}
knitr::include_graphics(paste(gridmaps.path,"SARD-PIL-[5,9].png",sep=''))
```
```{r, echo=FALSE, out.width="90%",fig.cap = "SARD-PIL-(9,13]"}
knitr::include_graphics(paste(gridmaps.path,"SARD-PIL-(9,13].png",sep=''))
```
```{r, echo=FALSE, out.width="90%",fig.cap = "SARD-PIL-(13,17]"}
knitr::include_graphics(paste(gridmaps.path,"SARD-PIL-(13,17].png",sep=''))
```
```{r, echo=FALSE, out.width="90%",fig.cap = "SARD-PIL-(17,21]"}
knitr::include_graphics(paste(gridmaps.path,"SARD-PIL-(17,21].png",sep=''))
```
```{r, echo=FALSE, out.width="90%",fig.cap = "SPRA-SPR-[5,7]"}
knitr::include_graphics(paste(gridmaps.path,"SPRA-SPR-[5,7].png",sep=''))
```
```{r, echo=FALSE, out.width="90%",fig.cap = "SPRA-SPR-(7,10]"}
knitr::include_graphics(paste(gridmaps.path,"SPRA-SPR-(7,10].png",sep=''))
```
```{r, echo=FALSE, out.width="90%",fig.cap = "SPRA-SPR-(10,12]"}
knitr::include_graphics(paste(gridmaps.path,"SPRA-SPR-(10,12].png",sep=''))
```
```{r, echo=FALSE, out.width="90%",fig.cap = "SPRA-SPR-(12,14.5]"}
knitr::include_graphics(paste(gridmaps.path,"SPRA-SPR-(12,14.5].png",sep=''))
```
## hydrology gridmaps
```{r, echo=FALSE, out.width="90%",fig.cap = "Bottom Temperature (°C)"}
gridmaps.path<-'G:/Tarek/Ifremer/Papiers/DEFIPEL_MFA_V2/Hydro/Rasters/plots/'
knitr::include_graphics(paste(gridmaps.path,"BotTemp.png",sep=''))
```
```{r, echo=FALSE, out.width="90%",fig.cap = "Chla"}
knitr::include_graphics(paste(gridmaps.path,"Chla.png",sep=''))
```
```{r, echo=FALSE, out.width="90%",fig.cap = "Deficit of potential energy"}
knitr::include_graphics(paste(gridmaps.path,"DefEpot.png",sep=''))
```
```{r, echo=FALSE, out.width="90%",fig.cap = "Height of equivalent freshwater depth (m) "}
knitr::include_graphics(paste(gridmaps.path,"Heq.png",sep=''))
```
```{r, echo=FALSE, out.width="90%",fig.cap = "Mixed layer maximum depth"}
knitr::include_graphics(paste(gridmaps.path,"ProfHmel.png",sep=''))
```
```{r, echo=FALSE, out.width="90%",fig.cap = "Depth of pycnoclin"}
knitr::include_graphics(paste(gridmaps.path,"ProfPycn.png",sep=''))
```
```{r, echo=FALSE, out.width="90%",fig.cap = "Surface temperature (°C)"}
knitr::include_graphics(paste(gridmaps.path,"SurfTemp.png",sep=''))
```
```{r, echo=FALSE, out.width="90%",fig.cap = " Surface salinity (psu)"}
knitr::include_graphics(paste(gridmaps.path,"SurfSal.png",sep=''))
```
```{r, echo=FALSE, out.width="90%",fig.cap = "Satellite surface temperature (°C)"}
knitr::include_graphics(paste(gridmaps.path,"SST_SAT.png",sep=''))
```
---
title: "Space-time distribution of clupeiforms habitats in the French Atlantic and Mediterranean areas."
subtitle: "DEFIPEL project, task 2.5 deliverable"
author: Mathieu Doray, Ifremer / EMH
date: "`r Sys.Date()`"
abstract: "The space-time distribution of clupeiforms habitats in the French Atlantic (Bay of Biscay) and Mediterranean (Gulf of Lion) areas are characterised, based on data collected during routine integrated acoustic surveys. Clupeiforms habitats have been characterised in the Bay of Biscay and Gulf of Lion, based on data collected in springtime during the PELGAS survey (2000-2019), and in summertime during the PELMED survey (2003-2019)."
header-includes:
- \usepackage{float}
- \floatplacement{figure}{H}
output:
bookdown::pdf_document2: default
bookdown::html_document2: default
# pdf_document:
# extra_dependencies: flafter
# html_document:
# df_print: paged
---
```{r setup, include=FALSE}
knitr::opts_chunk$set(echo = FALSE)
#,fig.pos = "!H", out.extra = "")
library(knitr)
library(bookdown)
pref='/media/mathieu/IfremerMData/'
pref='G:/'
path.export.rasters.SPF=paste(pref,'PELGAS/Data/gridMaps/Poissons/gridMapBBspSizeDepth/thrQuantile98/xycor/res0.25/Rasters/',sep='')
rasterMosaics.list=list.files(path.export.rasters.SPF,pattern='rasterMosaic*')
rasterMosaics.list.ane=rasterMosaics.list[substr(rasterMosaics.list,14,21)=='ENGR-ENC']
rasterMosaics.list.ane.caption=gsub('rasterMosaic_','',rasterMosaics.list.ane)
rasterMosaics.list.ane.caption=gsub('.png','',rasterMosaics.list.ane.caption)
rasterMosaics.list.ane.caption=gsub('[(]','',rasterMosaics.list.ane.caption)
rasterMosaics.list.ane.caption=gsub('[]]','cm',rasterMosaics.list.ane.caption)
rasterMosaics.list.ane.caption=gsub('[,]','-',rasterMosaics.list.ane.caption)
rasterMosaics.list.ane.caption=gsub('[_]','-',rasterMosaics.list.ane.caption)
rasterMosaics.list.pil=rasterMosaics.list[substr(rasterMosaics.list,14,21)=='SARD-PIL']
rasterMosaics.list.pil.caption=gsub('rasterMosaic_','',rasterMosaics.list.pil)
rasterMosaics.list.pil.caption=gsub('.png','',rasterMosaics.list.pil.caption)
rasterMosaics.list.pil.caption=gsub('[(]','',rasterMosaics.list.pil.caption)
rasterMosaics.list.pil.caption=gsub('[]]','cm',rasterMosaics.list.pil.caption)
rasterMosaics.list.pil.caption=gsub('[,]','-',rasterMosaics.list.pil.caption)
rasterMosaics.list.pil.caption=gsub('[_]','-',rasterMosaics.list.pil.caption)
rasterMosaics.list.spr=rasterMosaics.list[substr(rasterMosaics.list,14,21)=='SPRA-SPR']
rasterMosaics.list.spr.caption=gsub('rasterMosaic_','',rasterMosaics.list.spr)
rasterMosaics.list.spr.caption=gsub('.png','',rasterMosaics.list.spr.caption)
rasterMosaics.list.spr.caption=gsub('[(]','',rasterMosaics.list.spr.caption)
rasterMosaics.list.spr.caption=gsub('[]]','cm',rasterMosaics.list.spr.caption)
rasterMosaics.list.spr.caption=gsub('[,]','-',rasterMosaics.list.spr.caption)
rasterMosaics.list.spr.caption=gsub('[_]','-',rasterMosaics.list.spr.caption)
#knitr::knit_hooks$set(plot = function(x, options) {
# hook_plot_tex(x, options)
#})
```
# Introduction
- Objectives:
- analyse springtime distribution of clupeiforms in the Bay of Biscay (BoB) in space and time, based on survey (PELGAS) data
- describe main spatial patterns and temporal trends in clupeiforms distribution in springtime from 2000 to 2019
# Material and methods
- Data
- 0.25° gridmaps of 2000-2019 clupeiforms biomass (anchovy, sardine, sprat), per 5cm length class and depth stratum from PELGAS survey
- Space-time MFA (\~ grouped PCA) on series of clupeiforms biomass gridmaps (FactoMineR package)
- Rows: gridmaps cells
- Columns: Clupeiforms biomass per 5cm length class and depth stratum
- Groups: years
- MFA results interpreted in:
- MFA space: individuals (cells) / variables / groups (time)
- Geographical space: MFA maps
# Results
## Deviance explained by MFA axis
```{r fig1, echo=FALSE, out.width="50%",fig.cap='Deviance explained'}
path.export.results.clup=paste(pref,'PELGAS/Resultats/gridMaps/MFA/Poissons/2000-2019/Clupeiforms/',sep='')
load(file=paste(path.export.results.clup,"MFAclupResults.RData",sep=''))
print(paste(dim(big.fssd.clup.MFAs3)[1],'map cells x',dim(big.fssd.clup.MFAs3)[2],'variables',
'grouped into',length(stnnc2.clup),'years'))
NMFA95=dimnames(res.fssd.MFA.clup$eig[round(res.fssd.MFA.clup$eig[,3])>=95,])[[1]]
print(paste('95% deviance explained by',substr(NMFA95[1],6,7),'MFA components'))
barplot(res.fssd.MFA.clup$eig[,3],main="Cumulative % of variance explained",
names.arg=1:nrow(res.fssd.MFA.clup$eig))
```
## MFA individuals (cells) plane
Map cells in MFA1:3 planes are presented in Figure \@ref(fig:fig2):
```{r fig2, echo=FALSE, out.width="70%",message=FALSE,fig.cap="Map cells in MFA1:3 plane"}
library(factoextra)
library(gridExtra)
p1 = fviz_mfa_ind(res.fssd.MFA.clup, col.ind = "contrib",title='')
p2 = fviz_mfa_ind(res.fssd.MFA.clup, col.ind = "contrib",axes = c(2, 3),title='')
grid.arrange(p1, p2, ncol = 2)
```
## Variables contributions to MFA1
```{r fig3, echo=FALSE, out.width="100%",fig.cap = "Variables with abs(correlation) with MFA1 higher or equal to 0.6. Years in top pannel, species as rows, size categories in bottom pannel."}
knitr::include_graphics(paste(path.export.results.clup,'MFA1clupVarCor.png',sep=''))
```
Variables well correlated with MFA1 (Figure \@ref(fig:fig3)):
- MFA1 positive correlation with small/medium anchovy and sardine since 2002 and sprat since 2014
- MFA1 negative correlation with large sardine near sea surface from 2000 to 2002
## Variables contributions to MFA2
```{r fig4, echo=FALSE, out.width="100%",fig.cap = "Variables with abs(correlation) with MFA2 higher or equal to 0.6. Years in top pannel, species as rows, size categories in bottom pannel."}
knitr::include_graphics(paste(path.export.results.clup,'MFA2clupVarCor.png',sep=''))
```
Variables well correlated with MFA2 (Figure \@ref(fig:fig4)):
- MFA2 positive correlation with large anchovy near seabed from 2003 to 2010, negative correlation with small sprat at beginning and end of series.
## Clupeiforms mean spatial patterns: MFA maps
```{r fig5, echo=FALSE, out.width="90%",fig.cap = "MFA1 and 2 loadings mpaps."}
knitr::include_graphics(paste(path.export.results.clup,'rasterLevelPlot_MFAcoordGeoSpace.png',sep=''))
```
Two main spatial gradients in MFA1&2 loadings (Figure \@ref(fig:fig5)):
- MFA1: Gironde mouth vs. NW / offshore areas decreasing gradient
- MFA2: positive MFA2 path South of Belle Ile
## Clupeiforms spatial patterns inter-annual variability
```{r fig6, echo=FALSE, out.width="90%",fig.cap = "Inter-annual inertia of MFA1 and 2 loadings."}
knitr::include_graphics(paste(path.export.results.clup,'rasterLevelPlot_1-2_MFAcoordInertia.png',sep=''))
```
Inter-annual variability (inertia) higher in Gironde mouth (MFA1) and in high MFA2 patch (Figure \@ref(fig:fig6)).
## Global time trends in clupeiforms spatial distribution
```{r fig7, echo=FALSE, out.width="90%",fig.cap = "Clupeiforms MFA1 and 2 time series averaged over whole Bay of Biscay, 2000-2019."}
knitr::include_graphics(paste(path.export.results.clup,'fishMFA12timeSeries.png',sep=''))
```
No significant trend in clupeiforms MFA1&2 time series averaged over the whole Bay of Biscay.(Figure \@ref(fig:fig7)).
## Local clupeiforms time trends
### MFA1 time series in characteristic areas
```{r fig8, echo=FALSE, out.width="90%",fig.cap = "Mean clupeiforms MFA1 and 2 time series in Gironde characteristic area, 2000-2019."}
knitr::include_graphics(paste(path.export.results.clup,'MFA1clupTimeSeriesCharAreas.png',sep=''))