Prediction error through modelling concepts and uncertainty from basic data

Jansen, M.J.W.

doi:10.1007/978-94-017-3021-1_23

Cited by 22 publications

(32 citation statements)

References 8 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Thus, as measurement of SMN is expensive for farmers to carry out, model 2 is preferred as a practical model. Apart from practical considerations the model is best kept simple, because over-complication of models is not recommended 21 owing to the errors which are associated with the basic data and interpolation of eg rainfall data from`regions' to individual sites. …”

Section: Nitrogenmentioning

confidence: 99%

Estimation of the nitrogen requirement of winter wheat in the UK: a multiple regression approach

2002

View full text Add to dashboard Cite

In this study, multi-site crop response data were examined so as to improve guidelines for crop nitrogen requirement and ef®cient fertiliser use on winter wheat over a range of soil types. The parameters which had most individual effect on nitrogen requirement (the economic optimum amount of fertiliser N (Nopt)) were soil type and previous cropping. Interactions between soil type and soil N supply as measured by soil mineral nitrogen analysis were also important. Yield had no effect. Descriptive models were developed for Nopt. The simplest, based on N index, rainfall and soil type, accounted for 40% of the variation in Nopt. This was improved to 51% by adding spring analysis of soil mineral nitrogen. The model could be improved by inclusion of measured yield or a seasonal term, but the precision of predicted Nopt was not improved signi®cantly, 41% of cases being within AE25 kg ha À1 of the optimum without yield and 42% with yield. The study showed that some of the inputs currently used to determine N requirement need to be revised and additional terms added. However, even the best estimates have poor predictive precision. INTRODUCTIONAt present there is a link in fertiliser recommendations between yield and nitrogen (N) requirement (see eg Ref 1) which can lead to excessive applications of fertiliser. Recently we suggested 2 that many of the factors which affect yield, such as variety, date of sowing, take all (Gaeumannomyces graminis var tritici) and soil type, are unlikely to affect N requirement and that the yield adjustment may not be justi®ed. Formulation of fertiliser recommendations has been derived from extensive multi-site testing of responses in yield to applied N (see eg Ref 3). These experiments were analysed, and recommendations revised, successively rather than in toto. It is possible that comprehensive analysis of experimental series would provide improved precision in estimates of N requirements and an effective test of the yield effect. In this paper we describe an analysis of nine multi-site N experiments conducted by ADAS between 1981 and 1994 in order to identify and quantify those factors which have statistically signi®cant effects on N requirements, paying particular attention to expected yield.

show abstract

Section: Nitrogenmentioning

confidence: 99%

Estimation of the nitrogen requirement of winter wheat in the UK: a multiple regression approach

2002

View full text Add to dashboard Cite

show abstract

“…The problem was first recognised by ecologists (O'Neill, 1988;Meentemeyer, 1989), and recent awareness is growing in a number of other disciplines (Becker et al, 1999;Wagenet, 1998;Jansen and Stoorvogel, 1998). With one notable exception (Costanza, 1989), this growing interest in multi-scale analysis is not accompanied by special interest in multi-scale model validation, even though most modellers do not claim that their models are valid at the scale of analysis (Jansen, 1998). There is no reason to assume that the scale effect is less influential in model validation and the results of a multi-scale model could therefore potentially be misused if not coupled with a multi-scale validation.…”

Section: Scale Issuesmentioning

confidence: 99%

A method and application of multi-scale validation in spatial land use models

Kok

Farrow

Veldkamp

et al. 2001

Agriculture, Ecosystems & Environment

108

View full text Add to dashboard Cite

The majority of the large number of existing land use models lack a proper validation, often because of data problems. Moreover, despite recognition of the necessity to incorporate a multi-scale analysis, scale dependencies are normally not considered during validation. In this paper, a multi-scale land use change modelling framework, conversion of land use and its effects (CLUE), is calibrated for Costa Rica and validated at five spatial resolutions for Honduras and Costa Rica. Both countries experienced locally very strong actual land use changes. Calibration runs show that the model is very sensitive to changes in the autonomous development parameter, which defines the influence of the finest resolution. Validation results are very satisfactory for both countries. Especially, changes in major land use types are reproduced with the model. Changes in localised land use types are more difficult to project. The magnitude of gains and magnitude of losses are slightly underestimated in all cases. The multi-scale validation demonstrates that results improve strongly, and exponentially, with decreasing spatial resolution. Strong reduction of the number of observations results in a correlation between actual and modelled changes that approximates the perfect value of 1. The study demonstrates that the CLUE modelling framework can reproduce changes as they took place in Central America in the 1970s and 1980s, and shows how conclusions can differ depending on the scale at which validation is performed.

show abstract

“…If experimental data is not available then one must rely on values for typical autocorrelation lengths from the literature in order to do spatial interpolation (kriging). Describing uncertainty using geostatistics is not an activity exempt from uncertainty itself as variogram uncertainty may be large (Jansen, 1998) and spatial interpolation may be undertaken using different techniques.…”

Section: Geostatistics For Representing Spatial Variationmentioning

confidence: 99%

Uncertainty in soil physical data at river basin scale – a review

Keur¹,

Iversen²

2006

Hydrol. Earth Syst. Sci.

View full text Add to dashboard Cite

Abstract.For hydrological modelling studies at the river basin scale, decision makers need guidance in assessing the implications of uncertain data used by modellers as an input to modelling tools. Simulated solute transport through the unsaturated zone is associated with uncertainty due to spatial variability of soil hydraulic properties and derived hydraulic model parameters. In general for modelling studies at the river basin scale spatially available data at various scales must be aggregated to an appropriate scale. Estimating soil properties at unsampled points by means of geostatistical techniques require reliable information on the spatial structure of soil data. In this paper this information is assessed by reviewing current developments in the field of soil physical data uncertainty and adopting a classification system. Then spatial variability and structure is inspected by reviewing experimental work on determining spatial length scales for soil physical (and soil chemical) data. Available literature on spatial length scales for soil physical-and chemical properties is reviewed and their use in facilitating change of spatial support discussed. Uncertainty associated to the derivation of hydraulic properties from soil physical properties in this context is also discussed.

show abstract

Prediction error through modelling concepts and uncertainty from basic data

Cited by 22 publications

References 8 publications

Estimation of the nitrogen requirement of winter wheat in the UK: a multiple regression approach

Estimation of the nitrogen requirement of winter wheat in the UK: a multiple regression approach

A method and application of multi-scale validation in spatial land use models

Uncertainty in soil physical data at river basin scale – a review

Contact Info

Product

Resources

About