When Topoclimate were carrying out the temperature surveys in Southland, New Zealand we found that we continually had to justify the cost of the surveys to the farmers, the local community and to the Government Agencies and politicians.
A number of modelling scientists claimed that they could achieve the same results through the use of computer simulation models which would not involve such an intensive network of temperature dataloggers which were expensive to install and maintain.
We decided that the best way to resolve this debate would be to test both systems on a trial area and see which system gave the most credible result.
We chose the Five Rivers Plains area as our test case. The landscape of the trial Five Rivers Plains area is interesting and was selected as providing a good test of accuracy and detail for both systems. The Plains are surrounded by high mountains to the northwest and east and to a high range of hills in the south as shown in the three dimensional model of the area.
Note also that the Lumsden Long Term Weather Station is located just out of the area, so that its long term temperature data is relevant for both the modelling and measurement options. In the Topoclimate South Survey programme we located 122 dataloggers in this area to represent most thermal features of the landscape. There were nearly 50 different farms involved in this survey area, so the number of loggers per farm was only two to three depending on the complexity of landscapes encountered.
At first glance to the uninitiated, this may appear like overkill … and if we had adopted a solely modelling approach, we would have been very reliant on data from the Lumsden long Term weather Station based at Northern Southland Area High School on a terrace to the east of Lumsden.
The published map, below, shows GDDs at 4 degrees C as published by the Topoclimate South Trust based on the 122 logger network and detailed interpretation of the data on the ground by our Climatologist Bill Risk.
This map was independently audited by Professor Blair FitzHarris of Otago University and shows a number of significant microclimates of economic interest to the farmers who own land in this area.
Even with the detail involved in compiling this map, we considered that it is only about 98% accurate. You can observe from this map, with the logger sites overlaid, how the density of loggers was appropriate to the final map, and how the data from loggers contributed to the final pattern of GDD’s mapped for the Five Rivers District.
However, when we asked the Modellers to use a modelling approach for the same area (as was being promoted by other scientists) and used five data points only (actually more detailed than the measurements proposed by these scientists in their models) and the same modelling tools, such as an accurate digital terrain model, which accurately reproduces the contours of the land, they were only able to produce a map that, by my estimates, is about 5% accurate compared to the Topoclimate South Map in identifying the microclimates of the area.
Re-doing the modelling using the Topoclimate South density of data points (122) the Modellers were only able to generate a map that is about 75% accurate. The process required both detailed measurements and incorporation of local experience to get to the level of accuracy demanded by farmers for risk management decision making.
When the two output GDD4 maps are compared side-by-side, you can see that they have come closer to each other, because of the common number of logger data points used for each map. However, the modellers map is still considerably less accurate than the Topoclimate South map, because the Topoclimate South Staff incorporated local experience of climate patterns, such as local winds draining cooler air down gullies off the mountain slopes, which can’t be modelled with the same degree of accuracy.
In the end, the output maps are also about credibility. The local farmers in the Five Rivers Plain are more likely to act on, and use maps from data they have been involved in collation of, than from maps that are generated remotely from the comfort of a scientist’s desk.
As Topoclimate South Trust Chairman, Murray Ballantyne has stated, “the value in Topoclimate mapping is in the detail”.
For further information on the Monthly heat patterns in the Five Rivers Plains Area, please refer to the associated article: Monthly Heat Patterns in the Five Rivers Plains Area
