| Front page | | Contents | | Previous | | Next |

Establishment of a basis for administrative use of PestSurf

2 Models and scenarios

• 2.1 PestSurf models and scenarios
• 2.2 FOCUS models and scenarios - comparison to PestSurf

2.1 PestSurf models and scenarios

The PestSurf model and scenarios are documented in Styczen (2004a, 2004b and 2004c). PestSurf is based on two catchment models (for sandy and sandy loam conditions) that were developed and calibrated on measured data. The developed stream models were transformed into scenarios. For each catchment, a site with a natural water body was selected. A submodel was created and used as basis for a pond model that generated the water depth required by the DEPA. The ponds exist in two forms – a pond with macrophytes and an eutrophic pond with anaerobic sediment in the bottom.

The sandy loam catchment is 4.4 km² and the sandy catchment approximately 11.4 km². The layout of the stream system is shown in Figure 2.1 and Figure 2.2. The submodels for the sandy and sandy loam pond scenarios are 89 and 34 ha’s, respectively, but only part of the model area drains to the pond.

The pond is situated near the top end of Albjergbaek in the sandy loam catchment and north west of “Tilloeb” in the sandy catchment. In reality a small pond exists at that particular site in the sandy loam catchment, while the pond area in the sandy catchment is drained with ditches and receives groundwater from the western part of the catchment.

The processes included in PestSurf are wind drift, dry deposition, plant uptake, sorption, degradation and transport in the soil and groundwater, sorption to macrophytes and sediment in the stream as well as degradation, hydrolysis and photolysis. The effect of  emission/dry deposition is minimized because only the amount that  is deposited on the stream and buffer zones is allowed to leave the area. The rest is considered redeposited or substituted by deposition from surrounding areas.

The total agriculturally used area in the catchment is sprayed within half an hour when spraying occurs. The agricultural area makes up 82.3 % of the model area of the sandy loam catchment and 92.2 % of the sandy catchment. The rest is urban areas, roads or natural vegetation, which is never sprayed in the scenarios. For the pond scenarios, virtually 100 % of the surrounding area is sprayed (99.7 % for the sandy pond and 99.4 % for the sandy loam pond).

Click here to see Figure 2.1 and 2.2

Each of the streams are designated minimum buffer zones according to actual occurrence (Styczen et al., 2004b).  The natural buffer zones cover approximately 50% and 27% of the sandy and sandy loam catchment. The zones are summarized in Table 2.1.

Table 2.1. Buffer zones present along the sandy and sandy loam streams. The length of the stream is measured from the upstream end.
Tabel 2.1. Bufferzoner langs med det sandede og moræne-vandløbet. Længden af vandløbet er målt fra den opstrøms ende.

Climatic conditions and soil types are documented in Styczen (2004a). The precipitation is almost identical in the two scenario catchments, 852 mm/y and 854 mm/y for the sandy and sandy loam catchment, respectively. The simulation runs over 2*4 years. Calendar years were selected from the period 1990-1999. The first two years are selected to represent a very wet winter, the third year represents a very dry year and the fourth year is a normal year with respect to precipitation. Further, it has been attempted to secure that the precipitation in two spring and two autumn seasons are above average and in two are below average within the four year period. The water balance of the four year period is very close to the water balance for the 90’es for each of the sites.

One particular rainfall event included in the weather file for the sandy loam catchment in September 1994 and 1998 requires particular attention. On the 16^th and 17^th September, 56.6 mm and 18.5 mm of rain fall in this scenario. These values are found in other rainfall records from Funen as well, and are as such realisic. However, such values do not occur very often. For the 24-hour rainfall the value equals a 10year-occurrence and for the combined 48 hour-rainfall, it equals a one in 20-year event. This event is referred to in the report as the 20-year event.

The sandy loam soils contain macropores while the sandy soils do not. In case of the scenarios, the water flow is calculated for a number of different crops on the agricultural land and the pesticide simulations are carried out using the water file representing the relevant crop. The interactions between surface and groundwater differ somewhat in the two catchments and scenarios. In the sandy catchment, part of the flow leaves the catchment through sandy layers, crossing the catchment boundary. Only shallow groundwater may contribute to the stream via baseflow or drainage. Drainage occurs mainly from agricultural areas close to the stream, but the water table may be fed from areas further away. Some of the areas along the stream are meadows. No drainage is implemented for the sandy pond – it only receives contributions through inflowing groundwater and drift, while the sandy loam pond receives drainwater and has no permanent contact to the groundwater. It is clear from the simulations that the high groundwater level present in the vincinity of the sandy pond due to absence of drainage to some extent may be unrealistic in combination with intensive agricultural use. The consequence of the high groundwater level is a very fast transport of sprayed compounds to the groundwater and the pond.

The sandy loam catchment is intensively drained. Below the moraine clay is a sandy layer (an aquifer) receiving water from upstream of the catchment and delivering water over the downstream boundary. The stream cuts through the sandy layer at the downstream end. The effect of the sandy layer is that the downstream end of the sandy loam stream receives a baseflow contribution that more or less represents the whole catchment.

Earlier results have shown that the macropores implemented in the sandy loam catchment are too effective (too continuous and with too little interaction with the matrix), resulting in a higher contribution to groundwater than deemed realistic. An ongoing project develops a better description of macropore- and colloid transport.

Macrophytes are present in the streams and the ponds simulated. They are able to sorb and desorb pesticide.

During the project, an error related to the calculation of wind drift was discovered. The error resulted in a serious underestimation of the drift contribution, particularly for for the stream scenarios. All scenarios were re-run, meaning that two sets of simulations were produced. The correct calculation of drift and the assumption that the whole agricultural area is sprayed within 30 minutes leads to very high concentrations in the stream. In the discussion of results, some examples from the first set of simulations, where drift is considerably less important, will be used to exemplify how the results are expected to change if the mentioned assumption is changed.

2.2 FOCUS models and scenarios – comparison to PestSurf

Only three of the FOCUS SW-scenarios are of relevance here. These are the FOCUS SW-D4 stream and pond and the FOCUS SW-D3 ditch. The D4 scenario resembles the sandy loam catchment used in PestSurf with respect to rainfall, soils and crops. The D3-scenario is the FOCUS SW-scenario that resembles the PestSurf sandy catchment the most. However, only a ditch is represented, while PestSurf contains a pond and a stream. For all FOCUS SW-scenarios, the default values are used for data extraction and presentation. For the stream and pond it means that data are extracted 100 m down the stream or ditch. For the pond, the concentration is taken at the outlet.

As for PestSurf the precipitation is almost identical in the two scenarios. It is, however, somewhat lower in the FOCUS SW-scenarios than in the PestSurf scenarios, 693 mm/y for D3 and 692 mm/y for D4. The main reason for the difference is probably that no correction factors are used on the precipitation – present Danish recommendations lead to a yearly correction of precipitation amounts  of about 21 % (Allerup et al., 1998). In the FOCUS SW-scenarios, the waterbody simulation only covers about 1.3 year. The selected year is supposed to represent a 50 % leaching year.

The soil types for the FOCUS SW and PestSurf scenarios are listed in Table 2.2. The sandy soils in the FOCUS SW-scenarios are more sandy than the PestSurf sandy soils, but an important difference is that PestSurf contains a range of conditions rather than one single soil type. The sandy loam soils resemble each other, but again the variation in the PestSurf-conditions is rather large, particularly at depth.

The crops used in the two systems are very similar, and the same crops are used for simulations with the same compound.

In FOCUS SW, a minimum zone is defined between the sprayer and the water body [1 or 1.3 m for the ditch, 1.5 or 1.8 m for the stream and 3.5 or 3.8 m for the pond, depending on the crop. The same distance has been used in the PestSurf simulations. However, where natural buffer zones exist, they define the actual distance between the sprayer and the stream in PestSurf.

Table 2.2. Soil texture, carbon content and bulk density for soils used in FOCUS SW and PestSurf.
Tabel 2.2. Tekstur, carbonindhold og volumenvægt for jorde anvendt i FOCUS SW og PestSurf.

*not judged as representative, an average of 1.45 used for modelling (Styczen et al. 2004a)

The amount of spray drift should be comparable in the two models. FOCUS SW uses a 90^th percentile curve. In Pestsurf, the 95^th percentile curve presented by BBA (2000) is used to represent average drift conditions. This assumption is based on the fact that the wind speed in Denmark, even early in the morning where farmers normally spray, is about double the wind speed used for the drift measurements on which the curve is based. The same drift calculation is applied to each of several sprayings in a year in PestSurf. This is not the case in FOCUS SW, where the amount of drift per application decreases with increased number of applications. In Pestsurf, the drift is calculated, taking into account buffer zones and applied linearily over 30 min to each calculation point in the MIKE 11 model.

The spraying area differs considerably between some scenarios for the two models. For the pond, 100 % of the area is treated in the FOCUS SW and PestSurf-scenarios. The catchment to the FOCUS SW pond is 0.45 ha while the sub-models for the PestSurf ponds are 89 ha and 34 ha for the sandy and sandy loam catchment. In FOCUS SW infiltration is not allowed and the pond is regulated by a 1 m weir. In PestSurf infiltration is allowed, but outflow is regulated by a weir.

The FOCUS SW ditch is 100 m long and 1 m wide. It has an upstream area of 2 ha, which is not treated. The contribution of pesticide comes from a 1 ha field along the ditch, and input is received from spray drift and drainage. The water stems from drainage and baseflow is received from 20 ha, but the baseflow does not contain pesticide. The water depth in the ditch is regulated by a weir to minimum 0.3 m depth. PestSurf does not contain a ditch scenario. Small tributaries may resemble ditches, but the water depth is not regulated and they often run dry during the summer. They seldom reach a water depth of 30 cm. The FOCUS SW water bodies have rectangular cross sections while the PestSurf water bodies have triangular or trapezoid cross sections. For the streams a number of cross sections were measured and implemented. For the ponds, the cross sections are described in Styczen et al. (2004b).

The FOCUS SW stream is 100 m long and 1 m wide. It has an upstream catchment of 100 ha, of which only 20 % is treated with pesticide. 1 ha is treated with pesticide below the 100 ha catchment. Input is received from drainage and spray drift. Baseflow does not include pesticide. A minimum depth of 0.3 m is kept in the stream. As mentioned earlier, the PestSurf catchments are considerably larger (with stream lengths of about 4.2 km in the sandy catchment and 3.2 km in the sandy loam catchment) and the sprayed area is considerably larger than in the FOCUS SW-scenarios. Baseflow contains pesticide if the pesticide reaches the groundwater. The water depth in the stream is variable and much more dynamic than in the FOCUS SW-streams. When the sandy catchment is compared to the FOCUS D3-ditch, it should be noted that the flow rates in the stream are generally higher than in the ditch.

Macrophytes are not present in FOCUS SW.

The application time is treated differently in the two models. In PestSurf it can be specified whether spraying during rain is allowed. In the presented simulations it is not allowed. In this case, the time of spraying is moved forward until spraying is possible. It may, however, rain later in the day (sandy loam catchment) or the next day (sandy catchment). FOCUS SW has a special function that specifies the day of spraying within a month of the date specified in the user interface.

The calculations regarding diffusion into sediment differ somewhat between the two models. In the stream, Pestsurf operates with one sediment compartment, while FOCUS SW has a number of sediment layers. The diffusion coefficient used in the two modes is identical. For the ponds, both models operate with a number of sediment layers, but PestSurf uses an experimentally derived effective diffusion coefficient (Helweg et al., 2003) of 3.6 x 10^-2 m²/d compared to the value of 4.3 x 10^-5 m²/d used in FOCUS SW.

Other differences are summarised in Table 2.3.

Table 2.3. Differences between FOCUS SW and PestSurf not mentioned in the text.
Tabel 2.3. Forskelle mellem FOCUS SW og PestSurf, der ikke er nævnt i teksten.

| Front page | | Contents | | Previous | | Next | | Top |

Version 1.0 December 2006, © Danish Environmental Protection Agency