Silke Ullrich

Date of Award


Level of Access Assigned by Author

Open-Access Dissertation

Degree Name

Doctor of Philosophy (PhD)


Plant, Soil, and Environmental Sciences


Matt Liebman

Second Committee Member

Francis Drummond

Third Committee Member

Gregory Porter


Reduction of dependence on herbicides in potato production is important to limit environmental contamination and reduce production costs. An experiment was conducted from 1992 to 1997 in northern Maine to examine the effects of two rotation crops (a green-manure mixture of oat, pea and hairy vetch vs. barley) and two tillage practices (chisel vs. moldboard) on weed population dynamics in two-year potato rotations without the use of herbicides. Chenopodium album, Gnaphalium uliainosum, and a complex of cruciferous weeds (Brassica rapa, Brassica kaber, and Raphanus raphanistrum) were dominant. Oat-pea-vetch was more weed-suppressive than barley, however, crucifer density increased in both rotations. The increase was slower with moldboard- than with chisel-plowing. Only in the last year of the study after population buildup of the crucifer complex potato yield was reduced significantly by 27.8% due to the presence of weeds. This effect was probably enhanced by poor potato seed quality. A simulation model of the crucifer complex population dynamics with the same treatments as the field study was developed using parameters obtained from various independent experiments and validated using the experiment above. The model incorporates detail in the depth-structure of the seedbank, seedling emergence, and size-dependent seedling mortality due to cultivations in potato. Including weather in the model by categorizing growing seasons improved model performance. In 20-year simulations variation due to weather was greater than treatment differences. Sensitivity analysis identified seed survival near the soil surface, followed by seedling survival at hilling, seedling emergence in the first week after planting (WAP) and seed production in potato years as particularly influential on the crucifer seedbank. Simulations of contrasting weather conditions resulted in large differences in predicted population size and changes in ranking between treatments. Earlier hilling and a second hilling improved simulated weed control. Preemergence spring-tine harrowing reduced weed density more when performed 2 WAP than 1 WAP. Postemergence harrowing reduced weeds more 3 WAP than 4 WAP. To keep the crucifer population below damaging levels additional management practices need to be included in the cropping system. Further research should focus on the processes identified by sensitivity analysis and management practices acting upon them.