# Exploring management and environment effects on edge‐of‐field phosphorus losses with linear mixed models

## TL;DR

This study uses statistical models to explore how farming practices and environmental factors affect phosphorus loss from agricultural fields.

## Contribution

The study introduces linear mixed models to assess the combined effects of management and environmental variables on phosphorus loss.

## Key findings

- Manure application timing had only a marginal effect on phosphorus loss.
- No-till practices increased dissolved phosphorus loss, especially in frozen conditions.
- Higher soil phosphorus levels were linked to greater dissolved phosphorus loss.

## Abstract

Evaluating how weather, farm management, and soil conditions impact phosphorus (P) loss from agricultural sites is essential for improving our waterways in agricultural watersheds. In this study, rainfall characteristics, manure application timing, tillage, surface condition, and soil test phosphorus (STP) were analyzed to determine their effects on total phosphorus (TP) and dissolved phosphorus (DP) loss using 125 site‐years of runoff data collected by the University of Wisconsin Discovery Farms and Discovery Farms Minnesota. Three linear mixed models (LMMs) were then used to evaluate the influence of those factors on TP and DP losses: (1) a model that included all runoff events, (2) manured sites only, and (3) precipitation events only. Results show that the timing of manure application relative to the timing of a runoff event only had a marginal association with P loads and concentrations, although the majority of the runoff events were collected after 10 days of manure application. Tillage was as influential factor, with greater DP loads and concentrations associated with no‐till, especially during frozen conditions. Fields in this study had high STP values, but the model results only showed positive associations between DP load and DP flow‐weighted mean concentration (FWMC) loss at the 0‐ to 15‐cm depth. The precipitation event LMM (which included precipitation characteristics) was the model that resulted in the largest R
2 value. While the predictive capacity of the LMMs was low, they did illuminate the relative importance of management and environmental variables on P loss, and can be used to guide future research on P loss in this region.

Edge‐of‐field phosphorus losses are affected by more variables than can be included in a single controlled study.Linear mixed models can be used to evaluate the relative effects of management and environment on event P loss.Manure timing only had a marginal association with event P loss.Greater event‐dissolved P (DP) loss was associated with no‐till frozen conditions.Fields with greater soil test phosphorus (STP) were associated with greater event DP loss.

Edge‐of‐field phosphorus losses are affected by more variables than can be included in a single controlled study.

Linear mixed models can be used to evaluate the relative effects of management and environment on event P loss.

Manure timing only had a marginal association with event P loss.

Greater event‐dissolved P (DP) loss was associated with no‐till frozen conditions.

Fields with greater soil test phosphorus (STP) were associated with greater event DP loss.