Santiago Beguería

The output.hru file is one of the most important output files of the SWAT (Soil and Water Assessment Tool) simulation model. It contains daily or monthly (depending on how you configured your simulation) time series of about 70 variables that are computed within the HRUs (hydrologic response units) of your simulation. The HRUs are the basic simulation units of SWAT, so the output.hru file is about the most detailed output you can get from the model.

The output.hru file tends to be a huge text table, hard formatted with fixed columns (i.e., Fortran style). It mixes monthly and annual values of the output variables, and it is sometimes a pain to read. I wrote a little R function for reading SWAT output.hru files, and I thought it could be useful for somebody. It can be easily adapted to other SWAT output files or versions (I am still using SWAT 2005 here).

Usage: swat_readOutputhru(file,col=NULL,hru=NULL,year=NULL,lulc=NULL,ver=2012)

Arguments:

Result: a list object with monthly values stored as mon and annual values in anu.

Here’s the function:

swat_readOutputhru <- function(file,col=NULL,hru=NULL,year=NULL,lulc=NULL,ver=2012) {

	# format of the .hru file (SWAT 2012)
	fmt=list(var=c('LULC','HRU','GIS','SUB','MGT','MON','AREA','PRECIP','SNOFALL','SNOMELT','IRR',
		'PET','ET','SW_INIT','SW_END','PERC','GW_RCHG','DA_RCHG','REVAP','SA_IRR','DA_IRR','SA_ST',
		'DA_ST','SURQ_GEN','SURQ_CNT','TLOSS','LATQ','GW_Q','WYLD','DAILYCN','TMP_AV','TMP_MX',
		'TMP_MN','SOL_TMP','SOLAR','SYLD','USLE','N_APP','P_APP','NAUTO','PAUTO','NGRZ','PGRZ',
		'NCFRT','PCFRT','NRAIN','NFIX','F-MN','A-MN','A-SN','F-MP','AO-LP','L-AP','A-SP','DNIT',
		'NUP','PUP','ORGN','ORGP','SEDP','NSURQ','NLATQ','NO3L','NO3GW','SOLP','P_GW','W_STRS',
		'TMP_STRS','N_STRS','P_STRS','BIOM','LAI','YLD','BACTP','BACTLP',
		'WTAB','WTABELO','SNO_HRU','CMUP_KGH','CMTOT_KGH','QTILE','TNO3','LNO3','GW_Q_D','LATQ_CNT'),
		col=c(4,5,9,5,5,5,rep(10,79)))
	if (ver==2009) {
		fmt$var <- fmt$var[1:80]
		fmt$col <- fmt$col[1:80]
	}
	if (ver==2005) {
		fmt$var <- fmt$var[1:75]
		fmt$col <- fmt$col[1:75]
	}
	if (class(w)=='numeric') {
		col <- fmt$var[w]
	}

	# select columns
	if (!is.null(col)) {
		if (!('MON' %in% col)) {
			col <- c('MON',col)
		}
		if (!('LULC' %in% col)) {
			col <- c('LULC',col)
		}
		if (!('HRU' %in% col)) {
			col <- c('HRU',col)
		}
		w <- fmt$var %in% col
		fmt$var <- fmt$var[w]
		fmt$col <- ifelse(w,fmt$col,-fmt$col)
	}

	# read file, rearrange table
	res <- read.fwf(file,fmt$col,
		head=F,skip=9,encoding='latin1',
		strip.white=TRUE,nrow=-1,buffersize=20000)
	colnames(res) <- fmt$var
	res <- res[order(res$HRU),]

	# select hrus by number or by lulc
	if (!is.null(hru)) {
		res <- res[res$HRU>=min(hru) & res$HRU<=max(hru),]
	}
	if (!is.null(lulc)) {
		res <- res[res$LULC==lulc,]
	}

	# monthly and annual tables
	mon <- res[res$MON<=12,]
	anu <- res[res$MON>12,]
	colnames(anu) <- sub('MON','YEA',colnames(anu))
	w <- which(mon$HRU==mon$HRU[1] & mon$MON==mon$MON[1])
	ww <- c((w-1)[-1],nrow(mon))
	years <- min(anu$YEA):max(anu$YEA)
	mon$YEA <- NA
	for (i in 1:length(w)) {
		mon[w[i]:ww[i],][,'YEA'] <- years[i]
	}

	# select years
	if (!is.null(year)) {
		mon <- mon[mon$YEA>=min(year) & mon$YEA<=max(year),]
		anu <- anu[anu$YEA>=min(year) & anu$YEA<=max(year),]
	}
	
	# rearrange
	rownames(mon) <- rownames(anu) <- NULL
	w <- which(colnames(mon)=='MON')
	ww <- which(colnames(mon)=='YEA')
	mon <- mon[,c(colnames(mon)[c(1:w)],'YEA',colnames(mon)[-c(1:w,ww)])]

	# go
	return(list(mon=mon,anu=anu))
}

Example:

file <- './Scenarios/Default/TxtInOut/output.hru'
w <- c('PRECIP','SNOFALL','SNOMELT','IRR','ET',
    'SW_END','SA_ST','DA_ST','WYLD','REVAP')
# the following also works:
# w <- c(1,2,6,8,9,10,11,13,15,22,23,25,26,27,28,29)

# read the whole file
r <- swat_readOutputhru(file,col=w)

# read just one hru and one year
r <- swat_readOutputhru(file,col=w,hru=14,year=2009)

r$mon
r$anu