Modelling gene flow in poplars

(2005 – 2008) Philipps University of Marburg, Faculty of Biology, Department of Nature Conservation


Long-lived trees are able to impart genetic information to their descendants over many generations. In the case of genetically modified poplars, transgenes may enter natural stands of poplars in this manner (vertical gene transfer).

The aim was to develop a computer-based model to simulate gene flow between poplar populations in a real landscape. Data on the spread and outcrossing of poplars will be gathered from nature to test the model. The current project builds on preparatory work on modelling gene flow in trees.


  • A clone map has been prepared with the 61 most important types of hybrid poplars that are approved in Germany. With the genetic fingerprints deposited here it is possible to clearly allocate poplar clones to definite types.
  • A simulation model has been developed for the nature reserve Ederaue, with which the spread of transgenic traits in a real landscape can be simulated. In the model, every individual tree and its development can be simulated. In addition, the model can depict the genetic flow within and between species such as pollen or seed distribution.

Experiment description

Study area

The main study area was in an area of land near Fritzlar in Hessen measuring approximately 900 square kilometres, which has the Ederaue Nature Reserve at its centre. Within the nature reserve there is a native black poplar population of approximately 400 individual trees. More black poplars and also a number of hybrids capable of reproduction are growing within a 15 kilometre radius. Hybrids are commonly found on sports fields and paths and only rarely come from plantations. The Kühkopf Nature Reserve in Stockstadt am Rhein, which has a large number of black and hybrid poplars, serves as the control area.

The trees were mapped using GIS (geographical information system) data from regional forestry offices and GPS (global positioning system) data. Poplars are dioecious, having separate male and female plants. The gender of the trees was determined by the flower characteristics.

Black poplar (populus nigra)

Black poplar (Populus nigra), male

Catkin of the trembling poplar

Black poplar catkin, female

Photo: Marc Niggemann

Vitality analysis in dishes under different levels of cover

Ecological studies

Flowering phenology: The flowering periods of different poplar species (including black and hybrid poplars) must overlap in order for them to outcross. The aim was to determine as accurately as possible the duration and nature of the overlapping of the flowering periods, and the gender, of the species P. nigra and P. x canadensis.

Germination tests: To analyse the germination ability, seeds were collected from fertilised female poplars from different sites and from different species/hybrids. Since poplar seeds do not require a dormant period, they could be sown straight away. The number of germinating seeds was compared with the total number of seeds sown (germination percentage).

Vitality under competitive conditions. For these investigations the seed were grown in a greenhouse in planting dishes under varying canopy thicknesses. After four weeks the seedlings were harvested, measured and genetically classified according to species.

Molecular-biological studies

Analysis of gene flow via paternity testing: Paternity analysis should shed light on which male poplars were involved in fertilising the female poplars. Analysis of the data enabled the extent of pollen drift to be determined. This involved isolating DNA both from the seeds and from the seedlings of the harvested mother trees. This DNA was then tested for genetic markers (microsatellites), which provide information about a tree’s identity.

Compilation of a clone register: A clone register was developed to enable the rapid identification of the planted hybrid clones.

Development of the simulation model

Data collection: The “Introgression” program developed by the University of Marburg was being used for the computer simulation. Geographical, meteorological and genetic data were linked to model the gene flow as realistically as possible. Meteorological data was supplied by the weather stations and geographical data by the regional land survey offices and the forestry administration. The number and the distribution of the poplars in the area, and the germination capacity of the poplar seeds were collated in the investigated region (ecological data). The biomolecular examination to analyse the gene flow between individuals provided the genetic data. The analysis of paternity provided besides data on spatial distance also conclusions on the quality of the pollen.

Modelling risk scenarios: By producing scenarios, the model should indicate which sites are to be assessed as critical for growing transgenic poplar populations. The aim is to use the simulation program for other tree species and landscapes at a later date. Modelling is carried out in two stages: first the pollen dispersal is simulated and then this dispersal function can also be used for seed dispersal, with minor modifications to the parameters.


Distribution of poplars in the study area

Black and hybrid poplars existed in both areas under investigation. Since there were more black poplars in the Ederaue nature reserve than were originally known to the forestry commission, the study was focussed on this area. In total in both areas, 743 established trees and 380 young trees were calibrated by GPS; they were marked and leaf samples taken. Of the established trees, 566 were in the Ederaue reserve and 177 in the Kühkopf reserve. The 380 young trees are growing in a disused gravel pit in the western part of the Ederaue reserve.

Stages of flowering phenology:

Closed buds on a male poplar

closed buds on a male poplar

Open buds

Open buds

Pollen release

Pollen release

Receptive stigma on a female poplar

Receptive stigma on a female poplar

Photos: Marc Niggemann

Fig.1: Distribution of the flowering periods of trees at the Ederaue Nature Reserve.

The black and hybrid poplars flowered almost at the same time due to the late spring. The majority of male and female flowers did not overlap, but the odd male flower may have succeeded in pollinating the female flowers.

Harvesting seeds from tall poplars with the help of a Ruthmann aerial platform from the municipal utility services in Fritzlar.

Fig. 2: Distance of the father to the mother tree.
Grey bars = number of genetically proven fathers
Black bars = number of all potential fathers in the corresponding distance class

Fig. 3: The figure shows the simulated spatial pollen concentration of the two models. It is conspicuous that the pollen concentration is displaced to the Northeast in the meterological model (B) in comparison to the mathematical model (A). This result is a good indication of the regional wind data in the flowering period.

Ecological studies

Observing the flowering phenology: At the start of the growing period in 2006 the poplar populations in the main study area, the Ederaue Nature Reserve, and in the control area at Kühkopf, were visited twice a week to assess the extent of flowering. Male flowers have four distinct flowering stages and female flowers have six. The percentage of flowers at each stage was estimated.

In the Ederaue nature reserve 50 trees were examined; 17 were female and 21 male P. nigra trees, and 4 were female and 8 male P. x canadensis trees. Because of the late start of spring in 2006 the flowering periods of both species were almost the same (see Fig. 1).

In the Kühkopf nature reserve, 12 trees were examined in 2006; 1 female and 5 male P. x canadensis and 6 female P. nigra trees. It is not possible to determine the sex of the tree before the flowing stage, so that by chance no male P. nigra trees were chosen. The observation of the flowering stage in the reference area in 2006 showed similar results to those in the Ederaue. However, the overlapping of the flowering period between the two species was not so accurate as in the Ederaue.

Tests of germination capacity. In spring 2006 seeds were harvested from 8, and in spring 2007 from 11 mother trees. The time point was so chosen that the seed coats had just burst and the first wool could be seen. The seeds were put into petri dishes to germinate and counted after five and ten days. For both years, no clear difference in the germination rate for P. nigra and P. x canadensis could be determined. Almost all seeds germinated in both years (germination rate 90%).

Vitality under competitive conditions. For the seedlings grown under the three different grades of canopy, the size, number of leaves and the germination rates were determined. The genetic analysis of the seedlings showed that the black poplars generally acted more frequently as the father plant than the hybrid poplars.

Molecular biological examinations

Analysis of the genetic flow. As the isolation of DNA from the seed turned out to be impractical, all other studies were carried out using the seedlings.

Over 300 samples were taken from young trees on an area in the southwest region of the nature reserve. Analysis of the parentage showed that the majority of the parent trees were P. nigra trees in the immediate vicinity.

For the harvested seeds, most of the fathers were in the immediate vicinity of the mother. Only occasionally were fathers identified that grew at a distance of over 1000 metres (Fig. 2).

Assembling the clone map. The 61 most important hybrid poplars that have been approved in Germany were depicted in a clone map. The inspection and assessment of the map was performed by a reference laboratory. The work on this was finished in spring 2008.


Development of the model. Functions for seed distribution in the real landscape were incorporated into the computer program “Introgression”. A further adaptation allowed the programme to function with point orientation (previously a grid orientation) so that the construction of rows of trees and single trees was possible.

A critical aspect in the development of the model was the choice of a suitable dispersion function for the seed or pollen. For the poplar the dispersion is by wind, therefore it seemed obvious to gather data on the wind conditions. Therefore, for the years 2006 and 2007, a dispersion function was developed on the basis of meteorological dispersion models. Figure 3 shows the cloud dispersion for both simulation models used. The meteorological model shows the typical dispersion behaviour of wind-transported particles (Fig. 3). Data collected hourly from the German weather service on wind strength and direction, as well as the amount of sunshine and the natural cover of the landscape flowed into the calculations.

Application of the model. With the help of the model, differences in the fertility of the hybrids or the black poplars could be demonstrated and quantified.

In addition, various scenarios for the recolonisation of flood plains through natural crops could be examined using the simulation model. The findings still require further tests.