Nitrogen fertiliser trial database: a valuable resource. Climatic factors and first cut response to nitrogen application.
G.S. Rajendram, J. Waller, C. Cameron, R. Longhurst, J. Luo, A. Ghani and M. O’Connor. AgResearch, Ruakura Research Centre, East St, Private Bag 3123, Hamilton. Abstract Data from 1,272 nitrogen (N) fertiliser trials from around New Zealand over the last 80 years were collated onto an electronic database. Data collected included nitrogen (N) fertiliser forms, N application rates, plant dry matter (DM) yields, botanical composition, soil types and weather conditions (rainfall and air temperature). These data were sourced mostly from original trial records and reports. In this paper, a summary of the information gathered from the database and relationships between the first cut N response and climatic factors are presented. Nitrogen response was calculated as the extra daily kg of plant DM produced per kg of N applied. Most of the trials either used Nitrolime or Urea. A comparison of N responses between these two fertiliser products showed no significant overall difference. Late spring application gave the greatest and most reliable response to N application. In general, only weak relationships between climatic factors and response to N were established. The strongest relationship was found to be between response and average daily air temperature for each soil, under near-ideal conditions (average daily rainfall ≥ 1 mm and soil deficit ≤ 20 mm) where basal fertiliser had been applied. The relationship between N response and rainfall was weaker than the relationship between N response and temperature. An estimate of minimum average daily air temperature for an N response in spring was established as 5 0C and is consistent with overseas studies. This suggests that fertiliser should not be applied until the daily mean air temperature is > 5 0C. However, using data currently available, no single relationship has explained pasture response to applied N fertiliser application; perhaps not surprisingly. Better correlations may have been established for N response if more soil moisture or N status data were available and basal fertiliser application included more frequently in trial design. Only a few trials had soil moisture and N status data. Many trials, approximately 50%, did not have basal fertiliser applied. If any major or trace elements did limit pasture growth, this may have affected the magnitude of N responses. Some of the factors confirmed as influencing N response are temperature, soil type soil moisture and basal fertility. Therefore, a fertiliser decision support system needs to take into account at least these factors. This database represents a useful resource complementing future investigations. Introduction Nitrogen (N) fertiliser trials have been conducted in New Zealand over the past 80 years. The major organizations involved in N research have been MAF, MAFTech, and AgResearch. The DSIR also carried out several field trials on N during the 1970-80s. Approximately one-third of the N response trials conducted in NZ were carried out by O’Connor and his team at MAF between 1969-1975 (O’Connor, 1982). After MAFTech and the DSIR were disbanded, commercial fertiliser companies such as Ballance Agri-Nutrients and Ravensdown have provided the funding for most of the N research conducted since the mid-1990s until present, using technical expertise within the Crown Research Institutes and Universities to conduct trials. There have been numerous publications on the use of N fertiliser to pastures in the form of booklets and short articles, such as ‘N fertiliser use on pastures and crops’ (Ledgard, 1990’s); AgFact (Crush et.al., 1994) and ‘Fertiliser use on NZ dairy and sheep and beef farms’ (Roberts et. al., 1993). The findings from this study will help in part to develop ‘a decision support tree’. The purpose of the decision support tree is to ensure efficient N fertiliser use for desired environmental and production outcomes under New Zealand climate and soil conditions. The data obtained will also be useful for further improving the Overseer nutrient budget model. Method Field trials conducted throughout New Zealand have generally used the experimental designs and procedures outlined by Lynch (1966). First, data from all these field trials were obtained, collated, and entered into an electronic database. Climatic data from the nearest meteorological station were also included to assist the interpretation of the results. Data were later selected and statistical analysis undertaken. This paper presents information gained from the N trials and parameters used for setting up the database. The data were used to establish relationships between N response and climatic factors and a summary of this analysis is presented. Statistical Methods Much of the basic analysis was carried out in Excel, filtering the graphs according to various regimes: No summer months were used in the analysis of the data, as there were concerns over potential issues with sward composition. Trials where the control plot yields exceeded 5,000 kg DM/ha and where the days of growth did exceed 100 were excluded. There was a range of fertiliser products used. This study will consider only Nitrolime and Urea, the most common products by far. A discussion of the difference between the two is presented in the discussion section. Summaries of the regressions as well as the best models are provided in the Results and Discussion section. Note that these regressions are probably best described as ‘explanatory’ rather than ‘predictive’ models. Measure of Response Data were restricted only to first cut response for this study. In this report, the 50 kg N/ha rate is used to calculate the daily kg DM produced per kg of N applied. The measure of response considered in this study is “unit N”, which is defined as: ‘the increase in rate of growth per unit N applied’. Unit N = (Treatment Yield – Control Yield) / N Applied (kg) / Days of growth Yields are expressed as kg dry matter (DM). N applied is expressed in kg. Trials in the database did not always have a 50 kg N/ha rate. In order to include the vast majority of trials in the database for this report, application rates in the range 40-60 kg N applied were used. This requires that approximate linearity of the response be assumed over this range, a feature