--- title: "Getting started" output: rmarkdown::html_vignette vignette: > %\VignetteIndexEntry{Getting started} %\VignetteEngine{knitr::rmarkdown} %\VignetteEncoding{UTF-8} --- ```{r, include = FALSE} knitr::opts_chunk$set( collapse = TRUE, comment = "#>", warning = F, message = F, fig.align = "center" ) ``` ## Installation Install the package from [r-universe](http://tbep-tech.r-universe.dev/ui/#builds) as follows. The source code is available on [GitHub](https://github.com/tbep-tech/tbeploads). ```{r eval = F} # Install tbeploads in R: install.packages('tbeploads', repos = c('https://tbep-tech.r-universe.dev', 'https://cloud.r-project.org')) ``` Load the package in an R session after installation: ```{r} library(tbeploads) ``` ## Usage Load estimates are broadly defined as domestic point source (DPS), industrial point source (IPS), material losses (ML), nonpoint source (NPS), atmospheric deposition (AD), and groundwater sources and springs (GW). The functions are built around these sources with unique inputs for each. ### DPS The domestic point source (DPS) functions are designed to work with raw entity data provided by partners. The core function is `anlz_dps_facility()` that requires only a vector of file paths as input, where each path points to a file with monthly parameter concentration (mg/L) and flow data (million gallons per day). The data also describe whether the observations are end of pipe (direct inflow to the bay) or reuse (applied to the land), with each defined by outfall Ids typically noted as D-001, D-002, etc. and R-001, R-002, etc, respectively. Both are estimated as concentration times flow, whereas reuse includes an attenuation factor for land application depending on location. The file names must follow a specific convention, where metadata for each entity is found in the `facilities()` data object using information in the file name. For convenience, four example files are included with the package. These files represent actual entities and facilities, but the data have been randomized. The paths to these files are used as input to the function. Non-trivial data pre-processing and quality control is needed for each file and those included in the package are the correct format. The output is returned as tons per month for TN, TP, TSS, and BOD and million cubic meters per month for flow (hy). ```{r} dpsfls <- list.files(system.file('extdata/', package = 'tbeploads'), pattern = 'ps_dom', full.names = TRUE) anlz_dps_facility(dpsfls) ``` The `anlz_dps()` function uses `anlz_dps_facility()` to summarize the DPS results by location as facility (combines outfall data), entity (combines facility data), bay segment (combines entity data), and as all (combines bay segment data). The results can also be temporally summarized as monthly or annual totals. The location summary is defined by the `summ` argument and the temporal summary is defined by the `summtime` argument. The `fls` argument used by `anlz_dps_facility()` is also used by `anlz_dps()`. The output is tons per month for TN, TP, TSS, and BOD and as million cubic meters per month for flow (hy) if `summtime = 'month'` or tons per year for TN, TP, TSS, and BOD and million cubic meters per year for flow (hy) if `summtime = 'year'`. ```{r} # combine by entity and month anlz_dps(dpsfls, summ = 'entity', summtime = 'month') # combine by bay segment and year anlz_dps(dpsfls, summ = "segment", summtime = "year") ``` ### IPS The industrial point source (IPS) functions are designed to work with raw entity data provided by partners and are similar in functionality to the DPS functions. The core function is `anlz_ips_facility()` that requires only a vector of file paths as input, where each path points to a file with monthly parameter concentration (mg/L) and flow data (million gallons per day). Loads are estimated as concentration times flow. The file names must follow a specific convention, where metadata for each entity is found in the `facilities()` data object using information in the file name. For convenience, four example files are included with the package. These files represent actual entities and facilities, but the data have been randomized. The paths to these files are used as input to the function. As before, non-trivial data pre-processing and quality control is needed for each file and those included in the package are the correct format. The output is returned as tons per month for TN, TP, TSS, and BOD and million cubic meters per month for flow (hy). ```{r} ipsfls <- list.files(system.file('extdata/', package = 'tbeploads'), pattern = 'ps_ind_', full.names = TRUE) anlz_ips_facility(ipsfls) ``` The `anlz_ips()` function uses `anlz_ips_facility()` to summarize the IPS results by location as facility (combines outfall data), entity (combines facility data), bay segment (combines entity data), and as all (combines bay segment data). The results can also be temporally summarized as monthly or annual totals. The location summary is defined by the `summ` argument and the temporal summary is defined by the `summtime` argument. The `fls` argument used by `anlz_ips_facility()` is also used by `anlz_ips()`. The output is tons per month for TN, TP, TSS, and BOD and as million cubic meters per month for flow (hy) if `summtime = 'month'` or tons per year for TN, TP, TSS, and BOD and million cubic meters per year for flow (hy) if `summtime = 'year'`. ```{r} # combine by entity and month anlz_ips(ipsfls, summ = 'entity', summtime = 'month') # combine by bay segment and year anlz_ips(ipsfls, summ = "segment", summtime = "year") ``` ### ML Material losses (ML) are estimates of nutrient loads to the bay primarily from fertilizer shipping activities at ports. Historically, loadings from material losses were much higher than at present. Only a few entities report material losses, typically as a total for the year and only for total nitrogen. The material losses as tons/yr are estimated from the tons shipped using an agreed upon loss rate. Values reported in the example files represent the estimated loss as the total tons of N shipped each year multiplied by 0.0023 and divided by 2000. The total N shipped at a facility each year can be obtained using a simple back-calculation (multiply by 2000, divide by 0.0023). The core function is `anlz_ml_facility()` that requires only a vector of file paths as input, where each file should be one row per year per facility, where the row shows the total tons per year of total nitrogen loss. The file names must follow a specific convention, where metadata for each entity is found in the `facilities()` data object using information in the file name. For convenience, four example files are included with the package. These files represent actual entities and facilities, but the data have been randomized. The paths to these files are used as input to the function. The output is nearly identical to the input data since no load calculations are used, except results are shown as monthly load as the annual loss divided by 12. Additional empty columns (e.g., TP load, TSS load, etc.) are also returned for consistency of reporting with other loading sources. ```{r} mlfls <- list.files(system.file('extdata/', package = 'tbeploads'), pattern = 'ps_indml', full.names = TRUE) anlz_ml_facility(mlfls) ``` The `anlz_ml()` function uses `anlz_ml_facility()` to summarize the IPS results by location as facility, entity (combines facility data), bay segment (combines entity data), and as all (combines bay segment data). The results can also be temporally summarized as monthly or annual totals. The location summary is defined by the `summ` argument and the temporal summary is defined by the `summtime` argument. The `fls` argument used by `anlz_ml_facility()` is also used by `anlz_ml()`. The output is tons per month of TN if `summtime = 'month'` or tons per year of TN if `summtime = 'year'`. Columns for TP, TSS, BOD, and hydrologic load are also returned with zero load for consistency with other point source load calculation functions. Material loss loads are often combined with IPS loads for reporting. ```{r} # combine by entity and month anlz_ml(mlfls, summ = 'entity', summtime = 'month') # combine by bay segment and year anlz_ml(mlfls, summ = "segment", summtime = "year") ``` ### AD Loading from atmospheric deposition (AD) for bay segments in the Tampa Bay watershed are calculated using rainfall data from weather stations in the watershed and atmospheric concentration data from the Verna Wellfield site. Rainfall data must be obtained using the `util_ad_getrain()` function before calculating loads. For convenience, daily rainfall data from 2017 to 2023 at sites in the watershed are included with the package in the `ad_rain` object. ```{r} head(ad_rain) ``` The Verna Wellfield data must also be obtained from as monthly observations. This file is also included with the package and can be found using `system.file()` as follows: ```{r} vernafl <- system.file('extdata/verna-raw.csv', package = 'tbeploads') vernafl ``` During load calculation, the Verna data are converted to total nitrogen and total phosphorus from ammonium and nitrate concentration data using the `util_ad_prepverna()` function. Total nitrogen and phosphorus concentrations are estimated from ammonium and nitrate concentrations (mg/L) using the following relationships: $$ TN = NH_4^+ * 0.78 + NO_3^- * 0.23 $$ $$ TP = 0.01262 * TN + 0.00110 $$ The first equation corrects for the % of ions in ammonium and nitrate that are N, and the second is a regression relationship between TBADS TN and TP, applied to Verna. AD loads are estimated using the `anlz_ad()` function, where total hydrologic load by bay segment is calculated from the rain data and total nitrogen and phosphorus load is calculated by multiplying hydrologic load by the atmospheric deposition concentrations from the Verna data. Total hydrologic load for each bay segment is calculated using daily estimates of rainfall at NWIS NCDC sites in the watershed. This is done as a weighted mean of rainfall at the measured sites relative to grid locations in each bay segment. The weights are based on distance of the grid cells from the closest site as inverse distance squared. Total hydrologic load for a sub-watershed is then estimated by converting inches/month to m3/month using the area of each bay segment. The distance data and bay segment areas are contained in the \code{\link{ad_distance}} file included with the package. ```{r} head(ad_distance) ``` The total nitrogen and phosphorus loads are then estimated for each bay segment by multiplying the total hydrologic load by the total nitrogen and phosphorus concentrations in the Verna data. The loading calculations also include a wet/dry deposition conversion factor to account for differences in loading during the rainy and dry seasons. Using `anlz_ad()` to estimate AD load is done as follows, where `ad_rain` is the rain data and `vernafl` is the path to the Verna Wellfield data. ```{r} anlz_ad(ad_rain, vernafl) ``` Results can be summarized by segment, baywide, monthly, or annually using the `summ` and `summtime` arguments. By default, loads are returned monthly for each segment. Note that Boca Ciega Bay and Boca Ciega Bay South results are returned separately. Only Boca Ciega Bay South is used when estimating total bay loads.