Harms Phenology presentation June 16 2005

The High Plains Initiative for Integrated Phenology: Where we are today: 

The High Plains Initiative for Integrated Phenology: Where we are today Sherri Harms Jose Martinez University of Nebraska – Kearney June 16, 2005

Objectives for 2004/2005: 

Objectives for 2004/2005 Gather information & communicate with each other What do we already know? What do we want to know? Gather data – what is available? Study the data – what does it tell us?

Issues We Would Like to Address: 

Issues We Would Like to Address The break-out sessions will address these issues & the related discussion questions.

What data is available?: 

What data is available? Wheat flowering data for several sites since 1935 (Stephen Baenziger/Len Nelson) Temperature & precipitation data by site for entire US (time lengths vary) (NOAA; NDMC; & see NADSS) Horticultural flowering dates for 4 species, 1978-2004 (Lincoln, NE) (Richard Sutton) (many gaps) Soybean yield data (2000-2004 from several NE sites) (L. Nelson) Sorghum bloom data 1987-2004 (Lincoln, NE) (Jeff Perdersen) Wheat stem rust data 1922-1992 several sites (USDA Cereal Disease Lab) Oceanic climatic data (NOAA) Other?

Preliminary Studies: What does our data tell us?: 

Preliminary Studies: What does our data tell us? Trends Temperature trends Wheat flowering date trends Horticultural date trends Sorghum Blooming date trends Associations/Correlations Wheat flowering dates & temperature Wheat flowering dates & horticultural flowering dates Wheat flowering dates & sorghum flowering dates Precipitation (Local) & Oceanic (Global) Conditions Climatic change verification Have wheat horticultural varieties, and sorghum dates been occurring earlier? If so, can this be attributed to climatic change?

Trends: 

Trends

Temperature Trends: 

Temperature Trends High Plains Regional Climate Center (1905-2004)

Temperature Trends: 

Temperature Trends Lincoln Data (from CALMIT) 1887-1998 March 1- May 30 weekly data Minimum temperatures showed a slight negative trend For example:

Temperature Trends: 

Temperature Trends Lincoln Data (from CALMIT) 1887-1998 March 1- May 30 weekly data Maximum temperatures showed a slight positive trend For example:

Wheat Flowering Date Trends: 

Wheat Flowering Date Trends

Nationwide Kharkov Flowering Dates: 

Nationwide Kharkov Flowering Dates

Sorghum 50% Bloom Date Trends: 

Sorghum 50% Bloom Date Trends

Horticultural Flowering Date Trends: 

Horticultural Flowering Date Trends

Associations/Correlations: 

Associations/Correlations

Associations of wheat flowering dates to temperature: 

Associations of wheat flowering dates to temperature Initial Study UNL Agronomy Farm (East Campus) & Mead NE Research Farm Wheat flowering dates from 1935 - 1998 Two wheat varieties: Kharkov & Scout Lincoln daily minimum temperature 1935 – 1998 Decision Trees & Neural Networks Data mining methods

Association of wheat flowering dates to temperature: 

Association of wheat flowering dates to temperature Standardize all data into 7 categories (based on standard deviations) Use minimum temperatures for a pre-defined number of weeks that precede the approximate flowering date Experiment I: Ten weeks prior to the average flowering date. Experiment II: Twenty four weeks prior to the average flowering date. Build a model based on these temperatures Performance measure: percentage of time the model arrived at the actual value on test data (using cross validation)

Experiment 1 Results: 

Experiment 1 Results

Partial Decision Tree for the Kharkov wheat variety 10-week dataset: 

Partial Decision Tree for the Kharkov wheat variety 10-week dataset

Kharkov Flowering Dates Vs. Thermal Dates: 

Kharkov Flowering Dates Vs. Thermal Dates

Wheat vs. Sorghum: 

Wheat vs. Sorghum Correlation Coefficent .33

Relationships between Horticultural, Wheat & Sorghum Flowering Dates: 

Relationships between Horticultural, Wheat & Sorghum Flowering Dates Confidence = examples covered by the premise / covered by the consequence Lift =confidence/ the proportion of all examples covered by the consequence

Slide22: 

Climatic Associations to Precipitation in Nebraska T. Tadesse, 2002

Sample Relationships: 

Sample Relationships ed = extremely dry; sd = severely dry; md = moderately dry Tadesse & Harms

Predicting Phenological Development in Winter Wheat: 

Predicting Phenological Development in Winter Wheat Xue, Wiess, Baenziger, 2004 Streck, Weiss, Xue, Baenziger, 2003 Calculate daily development rate based on Temperature Vernalization response function Photoperiod Used to predict flowering dates with good accuracy (RMSE 5-6 days)

Climatic change verification: 

Climatic change verification

An interesting problem: 

An interesting problem Are there statistical or computer science analysis tools that can tie temporal data to singular events? In other words, how do we tie climate data to phenological data?

Wheat Flowering Date Trends: 

Wheat Flowering Date Trends

Wheat Flowering Dates & Temperature: 

Wheat Flowering Dates & Temperature

Summary of 2004/2005 Activities: 

Summary of 2004/2005 Activities Gathered information & communicated with each other Gathered data – limited and with holes Studied the data Several data sets indicate trends toward earlier maturation in plants – why? Worth further exploring – global warming?

Where do we go from here?: 

Where do we go from here? What other data do we want to tie together? Repository for data (Tied to the National Phenology Network) Decision Support System Such as the National Agricultural Decision Support System (NADSS) for USDA Risk Management Association Built-in analysis tools Data access tools that can be used by researchers, policy makers, educators