Read and Review Data

Goal= cleaned data frame with average percent coral cover per transect survey, with additional information needed for the master coral database.

# read in original csv file

rc_carib_substrate <- read_csv("sample_data/Reef_Check_Caribbean_Substrate.csv")

## Parsed with column specification:
## cols(
##   .default = col_character(),
##   `Longitude Degrees` = col_double(),
##   `Longitude Minutes` = col_double(),
##   `Longitude Seconds` = col_double(),
##   `Latitude Degrees` = col_double(),
##   `Latitude Minutes` = col_double(),
##   `Latitude Seconds` = col_double(),
##   Year = col_double(),
##   depth = col_double(),
##   total = col_double(),
##   `Errors?` = col_logical()
## )

## See spec(...) for full column specifications.

# clean column names- lowercase, no spaces
rc_carib_substrate <- janitor::clean_names(rc_carib_substrate)

# Look at number of sites and countries in this dataset

length(unique(rc_carib_substrate$reef_name))  #612 unique reefs

## [1] 612

length(unique(rc_carib_substrate$country))  #28 countries

## [1] 28

# substrate_codes
unique(rc_carib_substrate$substrate_code)

##  [1] "HC" "NI" "OT" "RB" "RC" "RK" "SC" "SD" "SI" "SP"

length(unique(rc_carib_substrate$substrate_code))  #10 types

## [1] 10

# Review a sample of data
rc_carib_substrate[1:10, ] %>% kableExtra::kable(caption = "Raw Reef Check Data") %>% 
    kable_styling(bootstrap_options = "condensed", full_width = F)

Raw Reef Check Data
reef_id	reef_name	longitude_degrees	longitude_minutes	longitude_seconds	longitude_cardinal_direction	latitude_degrees	latitude_minutes	latitude_seconds	latitude_cardinal_direction	year	date	depth	substrate_code	segment_code	total	errors	what_errors	country	state_province_island	city_town	ocean
16.22.16W.15.14.21N	Champagne Outer	61	22	16	W	15	14	21	N	2006	04-Mar-06	9.1	HC	S2	7	TRUE	did not complete S4 of substrate	Dominica	NA	NA	Atlantic
16.22.16W.15.14.21N	Champagne Outer	61	22	16	W	15	14	21	N	2006	04-Mar-06	9.1	HC	S1	5	TRUE	did not complete S4 of substrate	Dominica	NA	NA	Atlantic
16.22.16W.15.14.21N	Champagne Outer	61	22	16	W	15	14	21	N	2006	04-Mar-06	9.1	HC	S3	8	TRUE	did not complete S4 of substrate	Dominica	NA	NA	Atlantic
16.22.16W.15.14.21N	Champagne Outer	61	22	16	W	15	14	21	N	2006	04-Mar-06	9.1	NI	S1	0	TRUE	did not complete S4 of substrate	Dominica	NA	NA	Atlantic
16.22.16W.15.14.21N	Champagne Outer	61	22	16	W	15	14	21	N	2006	04-Mar-06	9.1	NI	S2	0	TRUE	did not complete S4 of substrate	Dominica	NA	NA	Atlantic
16.22.16W.15.14.21N	Champagne Outer	61	22	16	W	15	14	21	N	2006	04-Mar-06	9.1	NI	S3	0	TRUE	did not complete S4 of substrate	Dominica	NA	NA	Atlantic
16.22.16W.15.14.21N	Champagne Outer	61	22	16	W	15	14	21	N	2006	04-Mar-06	9.1	OT	S2	0	TRUE	did not complete S4 of substrate	Dominica	NA	NA	Atlantic
16.22.16W.15.14.21N	Champagne Outer	61	22	16	W	15	14	21	N	2006	04-Mar-06	9.1	OT	S3	0	TRUE	did not complete S4 of substrate	Dominica	NA	NA	Atlantic
16.22.16W.15.14.21N	Champagne Outer	61	22	16	W	15	14	21	N	2006	04-Mar-06	9.1	OT	S1	0	TRUE	did not complete S4 of substrate	Dominica	NA	NA	Atlantic
16.22.16W.15.14.21N	Champagne Outer	61	22	16	W	15	14	21	N	2006	04-Mar-06	9.1	RB	S1	15	TRUE	did not complete S4 of substrate	Dominica	NA	NA	Atlantic

Cleaning Reef Check Data - Caribbean

Cleaning steps:
- A: Calculate percent cover of each substrate code PER SEGMENT
  - 1- Group by reef id, year, data, segment code, substrate code
  - 2- Create the “perc_seg” variable that is the % cover of each substrate code per segment
- B: Calculate percent cover of each substrate code PER SURVEY
  - 1- Group by reef_id, year, date, and substrate code
  - 2- Create variable “perc_cov”, which is the average percent cover for the transect, based off the “perc_seg” column ; Create n_samples by counting the number of segment codes, based on the grouping
- C: Clean columns to bind later with master_coral
  - 1- Pick distinct columns to keep
  - 2- Remove cols don’t need
  - 3- Combine cover and sd data for substrate code- need to do this because can’t use spread on 2 cols
  - 4- Spread data so coral and macro have their own column
  - 5- Split the cover & sd data into 2 separate columns
  - 6- Unite deg_min_sec data into 1 col for lat and lon- convert later
  - 7- Rename and mutate some columns
  - 8- Convert lat and lon into correct format
  - 9- Final column type cleaning (numeric values, create 4 digit years)

# A: percent cover of each substrate code PER SEGMENT (ie out
# of 40 points)
rc_carib_substrate <- rc_carib_substrate %>% group_by(reef_id, 
    year, date, segment_code, substrate_code) %>% mutate(perc_seg = (total/40) * 
    100)

# Note, if try to do % by segment and by survey at same time,
# the values are the same because of the way group_by is
# done-- this is why i separated this process into 2 steps


# B: Percent cover of each substrate code PER SURVEY (out of
# 160 points, or average of all the segments for that survey)
# Note: 8/20/18- Found that for some locations, there was
# more than 1 survey done on a particular day. To keep
# surveys separate, need to also group by 'depth'
rc_carib_substrate <- rc_carib_substrate %>% group_by(reef_id, 
    year, date, substrate_code, depth) %>% mutate(perc_cov = round(mean(perc_seg), 
    2), sd_cov = round(sd(perc_seg), 2), n_samples = length(segment_code))  #this will tell me how many segments were completed!

# C: Clean columns to be in a good format to bind rows to
# master_coral dataframe

rc_carib_to_bind <- rc_carib_substrate %>% # 1 find distinct surveys- for reef check these are the
# unique variables
distinct(reef_id, date, depth, substrate_code, .keep_all = TRUE) %>% 
    
# 2 get rid of columns we no longer need
select(-segment_code, -total, -state_province_island, -city_town, 
    -perc_seg, -errors, -what_errors) %>% 
# 3 filter out the 2 substrate codes of interest, hard coral
# (HC) & algae (NI)
filter(substrate_code == "HC" | substrate_code == "NI") %>% 
# can't use spread on 2 cols, so combo % cov and sd,
# separated by _
unite(temp, perc_cov, sd_cov, sep = "_") %>% 
# 4 use spread to put the % cover of each substrate code into
# its own column
spread(substrate_code, temp) %>% 
# 5 break apart HC and NI data back into cover and sd
separate(HC, into = c("coral", "coral_std"), sep = "_", remove = TRUE) %>% 
    separate(NI, into = c("macro", "macro_std"), sep = "_", remove = TRUE) %>% 
    
# 6 Unite lat and lon col with deg_min_sec together
unite(lat, latitude_degrees, latitude_minutes, latitude_seconds, 
    sep = " ") %>% unite(lon, longitude_degrees, longitude_minutes, 
    longitude_seconds, sep = " ") %>% 
# 7 rename columns
rename(source_id = reef_id, depth_m = depth) %>% mutate(subregion = "caribbean", 
    method = "line_transect", orig_source = "reef_check", data_source = "reef_check", 
    t_length = 20) %>% separate(date, into = c("day", "month", 
    "year"), sep = "-", remove = TRUE)




# 8: Converting the lat and lon from deg_min_sec to dec_deg
# formats. Use measurements package to do the conversion:

rc_carib_to_bind$lat <- measurements::conv_unit(rc_carib_to_bind$lat, 
    "deg_min_sec", "dec_deg")

rc_carib_to_bind$lon <- measurements::conv_unit(rc_carib_to_bind$lon, 
    "deg_min_sec", "dec_deg")

# Make sure to put lat and lon as numeric
rc_carib_to_bind$lon <- as.numeric(rc_carib_to_bind$lon)
rc_carib_to_bind$lat <- as.numeric(rc_carib_to_bind$lat)

# 9: Turn year into 4 digits (note: this data is from
# 2006-2017 so we can just add 20 to the beginning of each
# element)

rc_carib_to_bind$year <- paste(20, rc_carib_to_bind$year, sep = "")
rc_carib_to_bind$day <- as.numeric(rc_carib_to_bind$day)
rc_carib_to_bind$year <- as.numeric(rc_carib_to_bind$year)

# make coral data numeric
rc_carib_to_bind$coral <- as.numeric(rc_carib_to_bind$coral)
rc_carib_to_bind$coral_std <- as.numeric(rc_carib_to_bind$coral_std)
rc_carib_to_bind$macro <- as.numeric(rc_carib_to_bind$macro)
rc_carib_to_bind$macro_std <- as.numeric(rc_carib_to_bind$macro_std)

Check Data

# Preview a sample of data:

rc_carib_to_bind[1:10, ] %>% arrange(reef_name) %>% kableExtra::kable(caption = "Coral and Macroalgae % Cover from Caribbean Reef Check Surveys") %>% 
    kable_styling(bootstrap_options = "condensed", full_width = F)

Coral and Macroalgae % Cover from Caribbean Reef Check Surveys
source_id	reef_name	lon	longitude_cardinal_direction	lat	latitude_cardinal_direction	depth_m	day	month	year	country	ocean	n_samples	coral	coral_std	macro	macro_std	subregion	method	orig_source	data_source	t_length
32.23.48.2W.3.50.5.01S	Buraco da Raquel	32.39672	W	3.834722	S	3.0	25	Sep	2008	Brazil	Atlantic	4	1.25	2.50	44.38	13.44	caribbean	line_transect	reef_check	reef_check	20
16.22.16W.15.14.21N	Champagne Outer	61.37111	W	15.239167	N	9.1	4	Mar	2006	Dominica	Atlantic	3	16.67	3.82	0.00	0.00	caribbean	line_transect	reef_check	reef_check	20
16.22.16W.15.14.21N	Champagne Outer	61.37111	W	15.239167	N	10.5	8	Apr	2006	Dominica	Atlantic	3	22.50	4.33	0.00	0.00	caribbean	line_transect	reef_check	reef_check	20
32.26.27.84W.3.50.38.57S	Laje Dois Irmaos Funda	32.44106	W	3.844056	S	21.0	9	Nov	2010	Brazil	Atlantic	4	20.62	8.98	36.25	7.77	caribbean	line_transect	reef_check	reef_check	20
32.26.30W.3.50.45S	Laje Dois Irmaos Funda	32.44167	W	3.845833	S	23.0	23	Nov	2009	Brazil	Atlantic	4	21.88	11.06	6.25	6.29	caribbean	line_transect	reef_check	reef_check	20
32.26.26.9W.3.50.49.09S	Laje Dois Irmaos Rasa	32.44081	W	3.846972	S	15.0	21	Sep	2008	Brazil	Atlantic	4	46.25	19.31	30.62	23.66	caribbean	line_transect	reef_check	reef_check	20
32.25.21.69W.3.52.9.17N	Sueste	32.42269	W	3.869222	S	3.0	13	Nov	2010	Brazil	Atlantic	4	0.62	1.25	59.38	12.48	caribbean	line_transect	reef_check	reef_check	20
32.25.21.77W.3.52.5.36S	Sueste	32.42272	W	3.868167	S	3.0	19	Sep	2008	Brazil	Atlantic	4	3.75	4.79	67.50	16.71	caribbean	line_transect	reef_check	reef_check	20
32.25.21.77W.3.52.5.36S	Sueste	32.42272	W	3.868167	S	1.6	20	Nov	2009	Brazil	Atlantic	4	3.75	4.79	58.12	10.87	caribbean	line_transect	reef_check	reef_check	20
32.25.30.5W.3.52.5.36S	Sueste Trilha	32.42514	W	3.868167	S	2.0	25	Sep	2008	Brazil	Atlantic	4	5.62	5.15	56.25	20.97	caribbean	line_transect	reef_check	reef_check	20

Example Data Cleaning: Reef Check

lkm

9/12/2019

Read and Review Data

Cleaning Reef Check Data - Caribbean

Check Data