R语言学习笔记--利用R绘制T–S diagram

T–S diagram（温盐图解）是研究海洋水团、海水混合时的一种图解，物理海洋学上经常用到，以温度（T）为纵坐标，盐度（S）为横坐标。绘制T–S diagram常用ODV软件，然而ODV实在太......。下面提供R语言绘制法。

#先加载包
    library(testthat); library(gsw);     library(oce)
    library(ggplot2);  library(stringr); library(dplyr); library(plyr)
#选择GSW；Oceanographic Analysis with R: The oceanographic community has a choice of two sets of formulae for calculating seawater properties: the UNESCO formulation, popularized in the 1980s, and the Gibbs-SeaWater (GSW) formulation, proposed in 2010.
    options(oceEOS = 'gsw')

#读取CTD数据（cnv格式）
    setwd("D:/R/Data/Ocean/ctd") #该文件夹共4个文件，"S1.cnv" "S2.cnv" "S3.cnv" "S4.cnv"
    temp <- list.files( pattern = ".cnv")    

下面提供两种绘制方法

方案一：oce包绘图

    ctds<- vector("list", length(temp))    
    for (i in 1: length(temp)) {
        Cast <- read.oce ( temp[i] )
    #str(Cast)  #查看数据结构, cnv文件由metadata、data、processingLog三部分构成。
    #读取文件名和经纬度信息；下面4行代码需根据cnv的不同格式作相应调整。
        Filename  <- str_replace(temp[i], '\\.cnv', '') #字符串替换
        Geo <- unlist  (str_split(Cast@metadata$header[20 : 21],pattern = ' '))
        Lon <- as.numeric(Geo[3]) + as.numeric(Geo[4])/60 + as.numeric(Geo[5])  / 3600
        Lat <- as.numeric(Geo[8]) + as.numeric(Geo[9])/60 + as.numeric(Geo[10]) / 3600     
    #读取数据
        ctds[[i]] <- as.ctd( salinity       = Cast@data$salinity, 
                             temperature    = Cast@data$temperature, 
                             pressure       = Cast@data$pressure, 
                             type           = "SBE", 
                             station        = Filename, 
                             longitude      = Lon, 
                             latitude       = Lat, 
                             deploymentType = "profile")
        rm( Filename, Geo, Lon, Lat)
     }
    #合并所有CTD数据
        Data0 <- as.section(ctds) 
        summary(Data0)

#绘图
    plotTS(Data0, nlevels = 9, grid = FALSE, bg = "transparent",
           pch = 16, cex = 0.5,
           col.rho = "lightgray", cex.rho = 1, lty.rho = 1)

方案二：ggplot2包绘图

#建立空白数据框
    Data1 <- data.frame(Station     = character(0),  
                        Lon         = numeric(0),    
                        Lat         = numeric(0),   
                        Pressure    = numeric(0), 
                        Temperature = numeric(0),    
                        Salinity    = numeric(0),      
                        SA          = numeric(0),    
                        CT          = numeric(0))

    for (i in 1: length(temp)) {        
        Cast <- read.oce (temp[i])
        #str(Cast) #查看数据结构, cnv文件由metadata、data、processingLog三部分构成。
    #读取文件名和经纬度信息；下面4行代码需根据cnv的不同格式作相应调整。
        Filename  <- str_replace(temp[i], '\\.cnv', '') #字符串替换
        Geo <- unlist  (str_split(Cast@metadata$header[20 : 21],pattern = ' '))
        Lon <- as.numeric(Geo[3]) + as.numeric(Geo[4])/60 + as.numeric(Geo[5])  / 3600
        Lat <- as.numeric(Geo[8]) + as.numeric(Geo[9])/60 + as.numeric(Geo[10]) / 3600
    #提取数据到data.frame
        data.tem <- data.frame(Station     = Filename, 
                               Lon         = Lon, 
                               Lat         = Lat, 
                               Pressure    = Cast@data$pressure,
                               Temperature = Cast@data$temperature, 
                               Salinity    = Cast@data$salinity)
    #计算Absolute Salinity和Conservative Temperature
        data.tem$SA <- gsw_SA_from_SP(data.tem$Salinity, data.tem$Pressure,  longitude = Lon, latitude = Lat)
        data.tem$CT <- gsw_CT_from_t (SA = data.tem$SA,  t = data.tem$Temperature,  p = data.tem$Pressure)
    #合并数据
        Data1 <- rbind.fill(Data1, data.tem)
        rm( Filename, Geo, Lon, Lat)
    }    

#构建σ0等密度线数据  
# make TS long table
    TS <- expand.grid(
            SA = seq( floor( min( Data1$SA )),     ceiling( max( Data1$SA )),     length.out = 100),
            CT = seq( floor( min( Data1$CT )) - 3, ceiling( max( Data1$CT )) + 3, length.out = 100)
          )  #为了显示最上和最下的等密度线，CT的范围选择+-3
    TS$Density <- gsw_rho(TS$SA, TS$CT , 0) - 1000    
#选择拟绘出的等密度线
    isopycnals <- subset(TS,
                         round(SA,1) == (ceiling(max(TS$SA))-0.8) & #等密度线加标注的位置的x轴坐标，数值0.8根据图片效果作相应调整
                         round(Density,1) %in% seq(min(round(TS$Density*2)/2), 
                                                   max(round(TS$Density*2)/2),
                                                   by = 1)) #选择plot将绘出的等密度线上的数据
    isopycnals$Density <- round(isopycnals$Density, 1) #保留一位小数
    isopycnals <- aggregate(CT ~ Density, isopycnals, mean) #相同Density的CT值求均值

#绘图
    p <- ggplot() +
          geom_contour(data =TS, aes(x = SA, y = CT, z = Density), col = "grey", linetype = "dashed",
                       breaks = seq(min(round(TS$Density*2)/2), max(round(TS$Density*2)/2), by = 1)) +
          geom_text(data = isopycnals, aes(x = 35.2, y = CT, label = Density),
                    hjust = "inward", vjust = 0, col = "grey60", angle = 15 ) +
          geom_point(data=Data1, aes(SA, CT, col = Station)) 
    p;

#细节完善
    p + scale_x_continuous(name = "Absolute Salinity [g/kg]",      limits = c(33.4, 35.2)) +
        scale_y_continuous(name = "Conservative Temperature [°C]", limits = c(0, 32),breaks = seq(5, 30, 5)) +
        theme_bw() +
        annotate(geom = "text", x = 33.7, y = 32, label = "20", hjust = "inward", vjust = 0, col = "grey60", angle = 15) +
        annotate(geom = "text", x = 35.1, y = 32, label = "21", hjust = "inward", vjust = 0, col = "grey60", angle = 15) +
         theme(text = element_text(size = 14), axis.text = element_text(size = 12),
              panel.grid = element_blank())       

两种方法比较，方案一简单，但用不同颜色标记不同站位比较困难；方案二颜色美观，但总体较繁琐（尤其是等密度线）。欢迎提出更好的解决办法。

主要参考资料：ggTS 、Kelley D E, 2018. Oceanographic Analysis with R