The influence of diversity in agricultural landscapes on pest abundance remains controversial despite several studies on the subject. For example, the ecological role of crop diversity in determining the source-sink dynamics of populations of polyphagous Heliothis spp. is complex, and varies at field, farm, region and national scales. Recently, the cropping structures within farms in Xinjiang, China, were adjusted due to changes in market conditions. The effects of these rapid changes in cropping structure at the landscape scale on the dynamics of Helicoverpa armigera have, however, not been well studied. This is particularly so for intensive agriculture in isolated oasis ecosystems, such as the irrigated cropping regions within the deserts of northern Xinjiang. The influence of a mosaic landscape on the population dynamics of H.armigera was measured using light traps in 16-17 farms covering 240 km² in 2007-2009. Light traps were placed in clusters on commercial farms which varied in the relative amounts of cotton, tomatoes, sugar beet, corn and wheat that they grew. Moths were collected and identified every morning from the beginning of May to the middle of August in each of the three years of the study. The average number of moths caught each day/ trap and each month/ trap were calculated for individual farms for each year. Farms were classified according to whether they used complex or simple cropping systems. For the purposes of this study, farms wherein the acreage of conventional cotton constituted < 50% of the cropping area were arbitrarily classed as complex, whilst farms where cotton made up >50% of the cropped area were classed as simple (i.e. with a more homogeneous pattern of crops compared to complex farms).
There were at least 3 generations of H. armigera per year, and as moth abundance was highest in the 3rd generation, there appeared to be a progressive build up in numbers through the season. The abundance of H. armigera moths was strongly correlated to the complexity of the agricultural landscape which influenced the source-sink dynamics of these pests amongst the different farms. The abundance of moths in complex cropping landscapes was 2-8 times larger than that in simple cropping landscapes.
The area of cotton (as a proportion of arable area) within the study region overall was 88% in 2007, 56% in 2008, and 35% in 2009.A consistent difference in Simpson's Reciprocal Index remained between the complex and simple agricultural systems throughout the 3 years. Furthermore, the numbers of moths captured annually per trap was correlated positively with Simpson's reciprocal diversity for three consecutive years. The annual abundance of H. armigera per trap had a positive relationship with the percentage of corn, processing tomato and wheat grown, and a negative relationship with the percentage of conventional cotton grown. Our study demonstrates that complex landscapes, with a variety of alternative plant hosts provided for H. armigera can support larger numbers of moths, compared with simpler (conventional cotton mostly) systems. These findings imply that diverse landscapes could increase the risk of a pest outbreak in non- Bt cropping regions, and delay the evolution of H. armigera resistant to Bt-crops in the Bt-crop belt. This study also suggests that an agricultural diversity index should be used for improving predictive modeling of pest abundance.