This report details a third piece of work in a series prepared for Bay of Plenty Regional Council.  It outlines recent refinements made to the ROTAN model and describes several scenarios of land use change and mitigation.  Findings are intended to help managers develop policy by estimating the extent of export reduction required to meet the lake target of 435 tN/yr and how quickly the load to the lake is likely to respond to such reductions.  The results will be used by the University of Waikato to predict likely changes in lake water quality.

Previous reports provided the following:

  • The first report provides a description of fitting ROTAN to the observed daily flows in the nine major streams that flow into Lake Rotorua and to the observed lake outflow in the Ohau Channel
  • The second report describes fitting ROTAN to measured total nitrogen (TN) concentrations in the major streams and published estimates of nitrogen input to the lake.


This report is focused on the required reductions in diffuse sources of nitrogen in the catchment, which are largely on agricultural land.  This report does not discuss possible reductions in point source loads – these are included in each of the ROTAN models but do not vary between scenarios. This report provides the following:

  • Inform BoPRC managers about what can be achieved in terms of reducing the load of nitrogen entering Lake Rotorua through changing land use or changing the way land is managed
  • Quantify the total area of each current land use that will need to change to meet the lake load target
  • Indicate where in the catchment these land uses currently occur
  • Estimate the likely rate at which the nitrogen load to the lake will decrease after changes are made
  • Help develop effective policy regarding land use and land management


The report does not set out to identify parcels of land where it would be best to effect change in land use in or to reach the target in the least time and/or at least cost – although ROTAN has the potential to be used for that purpose in the future.



Input data included:

  • Surface catchments and aquifers
  • Land use maps
  • Land use categories
  • Spatial distribution of land use
  • Nitrogen yield from pasture
  • Nitrogen exports during development
  • Nitrogen delivery pathways
  • Point source contribution
  • Groundwater age


In this report, new input data has been introduced to the original model and changes were made to the model to facilitate long model runs and the model was recalibrated.  A new version of the model called ROTAN-1 was the result.  Some of the original model problems remained.  These included matching groundwater nitrogen loads owing to uncertainties in the extent of aquifer boundaries, groundwater age, land use patterns and nitrogen export rates.  However, the model fit was deemed sufficiently good for scenario modeling.  Several alternative versions of the ROTAN model were developed to test the sensitivity of predictions to uncertainties in key model coefficients and input data.  These versions were named ROTAN-2 to ROTAN-9.  While they were not calibrated as carefully as ROTAN-1 they provided valuable insights into model behaviour and reliability.



This report presents four scenarios of possible land use change (R-0, R-250, R-300 and R-350), aimed at achieving the target nitrogen load to the lake.  The focus is on predicting the magnitude (load reduction) and timing (response time) of changes in lake load.

The four scenarios are:

  • R-0 in which land use and nitrogen exports remain at their current levels from 2015-2100 (a “holding” scenario)
  • R-250 in which total nitrogen export from land is reduced by 250 tN/yr through a combination of land use change and a reduction in nitrogen yields
  • R-300 in which total nitrogen export from land is reduced by 300 tN/yr through a combination of land use change and a reduction in nitrogen yields
  • R-350 in which total nitrogen export from land is reduced by 350 tN/yr through a combination of land use change and a reduction in nitrogen export.

 Rotan graph


 ROTAN-1 provides a satisfactory match to key features of observed flows and concentrations

  • Some questions remain about the uniqueness of model calibration and hence the robustness of ROTAN-1 for making predictions
  • ROTAN-1 predictions are considered to be sufficiently reliable to inform policy and management
  • Predicted load reductions are considered to be robust but predicted response times are likely to have a higher uncertainty
  • Follow up modeling can be undertaken if more refined simulations are required

Associated Models

  • Rotorua Taupo Nitrogen Model (ROTAN)

    ROTAN is a GIS-based catchment hydrology and water quality model that models the effects of land-use change and nitrogen mitigation measures such as riparian margins and plant uptake on nitrogen export to waterways.

    More info


Prediction of Nitrogen loads to Lake Rotorua using the ROTAN model

Rutherford, K; Palliser, C; Wadwha, S.

Prepared by NIWA for Bay of Plenty Regional Council, 2011