Informations et ressources scientifiques
sur le développement des zones arides et semi-arides

Accueil du site → Doctorat → États-Unis → 1987 → The development of a crop specific drought index for winter wheat

University of Nebraska - Lincoln (1987)

The development of a crop specific drought index for winter wheat

Booysen Johann

Titre : The development of a crop specific drought index for winter wheat

Auteur : Booysen Johann

Université de soutenance : University of Nebraska - Lincoln

Grade : Doctor of Philosophy (Ph.D.) 1987

The purpose of this study was to evaluate the utility of wheat growth simulation models in deriving a crop specific drought index (CSDI). This new index features both the environment and the crop : transpiration ; root absorption ; phenological growth stages ; and the climatic and soil moisture variables. Four models were chosen to predict plant features from environmental input data.^ Environmental and physiological data were available from the 1985 and 1986 growing seasons for winter wheat at five locations in North America ranging from Lethbridge, Canada, to Lubbock, Texas. The potential for using models to monitor drought on a regional scale was evaluated under conditions of limited input data.^ Model performance was compared by examining statistical indices. Measured and model-predicted values of leaf area index (LAI), tiller number, soil moisture content at all observed depths, and final yield were compared. CERES, the selected model, was then tested using a twenty-two (22) year data set of crop production and meteorological conditions at North Platte, Nebraska. Comparisons were drawn between CSDI, Palmer Drought Severity Index (PDSI), and reported crop yields at North Platte.^ Several model performance statistics were included in a three dimensional graphical presentation to make comparison between models relatively easy. Acceptable performance was obtained from the soil moisture analysis and the CERES model did remarkably well in leaf area index estimation. However, the performance of all models on estimated tiller numbers was unsatisfactory. The performance of some models relative to yield estimates over all locations is somewhat inflated due to clustering of the data.^ Coefficients of determination obtained from the 22 year analysis at North Platte were low. Without proper historical records, however, it is not possible to improve on what is perhaps the weakest link, the accuracy of inputs. The CSDI showed definite advantages over the PDSI when compared to crop performance. It was shown that small gains in model accuracy will result in large gains in CSDI accuracy. The CSDI concept can be applied to other crops via other models and therefore has greater potential for improvement and use in drought monitoring than the PDSI or Crop Moisture Index.


Page publiée le 1er janvier 2015, mise à jour le 28 décembre 2016