Upload
others
View
1
Download
0
Embed Size (px)
Citation preview
RESULTS Four coun(es along the transi(on illustrate the reduced crop yield along the plant survival gradient. Corn yield was reduced by the drought for more years in NE than in IA similar to the spa(al pa?ern of na(ve grass survival. Line colors correspond to highlighted transi(on zone coun(es in figure 1.
The ecology of catastrophic events: understanding abrupt spa<al transi<ons in suscep<bility of grasslands and croplands to mul<-‐year drought
N. Dylan Burruss1, Debra P.C. Peters1,2, Jin Yao1, Kris M. Havstad1,2 and Stacey L.P. Scroggs1
1Jornada Basin LTER, NMSU, Las Cruces, NM, 2Jornada Experimental Range, USDA Agricultural Research Service, Las Cruces, NM, 3New Mexico State University ([email protected])
Period N PPT MAT PAW
(20-‐50cm) PAW
(0-‐20cm) Intercept R-‐
Square Pr > F Pre-‐Drought 41 0.05 -‐6.89 0.65 -‐ 39.44 0.84 <0.0001 Drought 41 0.10 -‐8.28 0.71 -‐ 12.66 0.94 <0.0001 Post-‐Drought 41 0.06 -‐8.61 1.19 -‐1.07 64.36 0.91 <0.0001
Significant regression coefficients for best models explaining the spa7al pa8ern in corn yield along the IA-‐NE transi7on zone.
During the drought, increases in grass mortality were related to decreases in precipita7on, increases in temperature and decreases in plant available water.
Related presenta<on: Peters et al. Can we use the past as a lens to the future? Using historic events to predict regional grassland and shrubland responses to mul(-‐year drought or wet periods under climate change. In session COS5 Climate Change I @ Mon Aug 10, 2:50 – 3:10 pm, room 319.
INTRODUCTION Much of the central grasslands region (CGR) of North America experienced a mul(-‐year extreme drought in the 1930s that combined with land management prac(ces to result in broad-‐scale plant mortality, massive dust storms, and losses of soil and nutrients.
RESEARCH OBJECTIVES 1. Compare the abrupt changes in na(ve grasses during the 1930’s drought along
the Iowa (IA) to Nebraska (NE) transi(on from reduc(ons in cover to high plant mortality using crop yield data through (me.
2. Determine the clima(c, soil, and land use drivers of these abrupt changes.
During the 1930s drought, all grassland types in the CGR were affected, but to varying degrees. In some areas, for example, na(ve grassland sites in southeastern Nebraska suffered high plant mortality while sites on similar soils located < 100 km to the east in southwestern Iowa only had reduc(ons in cover. The transi(on between severely impacted grasslands and not impacted grassland occurred over short distances, and our understanding of these abrupt
METHODS We first compared the spa(al pa?ern in corn yield to that of na(ve grass survivorship in the transi(on zone. Then we used stepwise regression to select the explanatory variables that best explain the spa(al varia(on in corn yield. We carried out the regression analysis separately for three periods: pre-‐drought (1926-‐32), drought (1933-‐’40) and post-‐drought (1941-‐’48). All sta(s(cal test were performed in SAS v. 9.4 with an alpha = 0.05. STUDY AREA: • 41 coun(es along the IA-‐NE transi(on zone. RESPONSE VARIABLES: • Annual corn yield for pre-‐drought, drought and post-‐drought periods. EXPLANATORY VARIABLES: • Period mean annual precipita(on, • Period mean annual temperature, • Plant available water at depths of 0-‐20cm and 20-‐50cm.
CONCLUSIONS 1. Difference in corn yield between IA and NE during the 1930s corn
produc(on suggest that corn yield is a good proxy for spa(al pa?erns in na(ve grasses during the 1930s.
2. Pa?erns in corn yield were related to PPT, MAT and PAW sugges(ng that the same drivers affected grass survivorship.
3. These cumula(ve effects of mul(-‐year drought challenge our ability to predict ecosystem responses to catastrophic events without sufficient long-‐term data.
• Long term corn yield chart was generated using the updated NASS data viewer at: h:p://innova<onchallenge.azurewebsites.net
Figure 1. Study coun7es occurring along the grassland transi7on zone.
Transi<on Zone Coun<es
Greene, IA
Fillmore, NE
Plant Available Water (mm)
Mean Temperature (°C) Mean Precipita<on (mm)
Temporal trends in Palmer drought severity index (PDSI) and data on grass survivorship during 1934.
Transi<on Zone Coun<es
transi(ons remains limited. Furthermore, sufficient long term ecological data is unavailable for this region and temporal period. Instead, Corn yield through (me from USDA long term agricultural data was used as a proxy for na(ve grass produc(on during the 1920’s and ‘30s.
17 -‐ 161
Annual Corn Harvest Value by years
Yield (BU/Acre) Area Planted (Acres) Area Harvest (Acres)
The drought period difference between acres planted and acres harvested was much greater in NE than in IA during the 1930s.
Nebraska
Iowa
• The delinea<on of the grassland types were mainly based on the map of level-‐3 EPA ecoregions. PDSI data source: Na<onal Clima<c Data Center (now Na<onal Centers for Environmental Informa<on) , NOAA. (h:ps://www.ncdc.noaa.gov/sotc/drought/201506)
• Grassland Survivorship published in Albertson, J.E. and Weaver F.W. (1936). Effects on the Great Drought on the Prairies of Iowa, Nebraska, and Kansas. Ecology, 17(4): 567-‐639.
• Long term agricultural data was obtained from the USDA Census of Agriculture for cropland and downloaded at h:p://quickstats.nass.usda.gov/
• Clima<c data source was Prism downloaded from h:p://www.prism.oregonstate.edu/. • Plant available water (=available water storage) at 0-‐5 cm soil depth . Data source: USDA Natural Resources Conserva<on Service
gSSURGO database (h:p://www.nrcs.usda.gov/wps/portal/nrcs/detail/soils/survey/geo/?cid=nrcs142p2_053628)
PDSI
PDSI
PDSI
PDSI
PDSI
Weaver Plant Survival
Grassland Regions