1. Osvaldo Sala

2. Projected Precipitation Change 1970-99 vs 2071-99 US National Climate Assessment 2014

3. Precipitation Variability is projected to increase IPCC. 2013. Climate Change 2013: The physical science basis. Contribution of working group I to the fifth assessment report of the Intergovernmental Panel on Climate Change. Cambridge, United Kingdom and New York, NY, USA.

4.  Grassland Shrubland  Hypothesis 2 Directional changes in water availability that favor grasses over shrubs or shrubs over grasses are reinforced through time.

5. Our scientific approach Observations Experiments Data Mining Modelling

6. + Peters et al (2011)

7. +

8. 2014 ambient + 80% - 80%

9. Solar panel Batter y Pump Intermediary tank (55 gal.) Float switch Interception plot Irrigation plot Filter ARMS automated rainfall manipulation system (Plots trenched to 40-60 cm) Gherardi and Sala, Ecosphere: 2013

11. Assess ecosystem sensitivity to precipitation Not to replicate climate change scenarios

12.  CC Impact = ʄ (Δ Climate, Ecosystem Sensitivity)  PPT Imp = ʄ (Δ PPT, Ecosystem Sensitivity to PPT)

13. Total Aboveground Net Primary Production + 80% Ambient - 80 %

14. Grass Aboveground Net Primary Production + 80% Ambient - 80 %

15. Shrub Aboveground Net Primary Production Ambient + 80 % - 80 %

16. Direct and Indirect Effects of Precipitation

17. Plant-Species Diversity H’ + 80% Ambient - 80 %

18.  There was an effect of time on ecosystem response variables to long-term changes in PPT  The effect of time varied for different response variables  Asymmetry Hypothesis – The absolute magnitude of the effect was different for increasing or decreasing PPT

20. 10 reps * 5 treat = 50 plots (2.5 x 2.5 m) Trenched 60 cm deep 20 rainout shelters 20 irrigated plots 10 control plots Methods

22. Effect of PPT Variability Gherardi & Sala PNAS 2015

23. The Mechanism Gherardi & Sala PNAS 2015

24. How do we explain these responses? Sala, Gherardi, Peters Climatic Change 2015 Modelling

25. Gherardi & Sala PNAS 2015 Effect of PPT Variance Increases through Time

26. Demise of grasses under high PPT variability favors shrubs

27.  Further explore the existence of thresholds ◦ Cumulative endogenous ◦ Stochastic exogenous ◦ Interaction between endogenous and exogenous  Mechanisms for indirect effects

28. Thank you Laureano Gherardi, Lara Reichmann Courtney Currier Kelsey Duffy Owen McKenna Josh Haussler

31. Collins et al 2011

32. Smith MD et al (2009)

33. a) Ecosystem response variables are proportional to water availability Response variable = b 0 + b 1 *PPT 0 + +

34. b) Ecosystem response variables are proportional to changes in water and to the time that the ecosystem has been exposed to the new condition Response variable = b 0 + b 1 *PPT + b 2 *Time + + - 0

35. 0 + + Peters et al (2011)

36. 0 + +

37. c) Acclimation / exhausting of resources Response variable = b 0 + b 1 *PPT + b 2 *Time + b 3 *Time*PPT 0 + + -

38.  The effect of time is asymmetric for reduced water and increased water 0 + + -

39.  The effect of time varies for different response variables 0 + + -

40. Solar panel Batter y Pump Intermediary tank (55 gal.) Float switch Interception plot Irrigation plot Filter ARMS automated rainfall manipulation system (Plots trenched to 40-60 cm) Gherardi and Sala, Ecosphere: 2013

41. Total ANPP Shrub ANPP Grass ANPP Spp Richness Diversity PPT

42.  Rejected H1a. There was an effect of time on ecosystem response variables to long-term changes in PPT, due to legacies in the ecosystem response.  Asymmetry – The absolute magnitude of the effect was different for increasing or decreasing PPT, i.e. spp loss with drought – no spp change with increased PPT  The effect of time varied for different response variables; may depend of the number of actors involved or the flow size relative to the pool size

43. Thank you Laureano Gherardi Lara Reichmann Owen McKenna Josh Haussler Kelsey Duffy Jose Anadon

44. NSF-Division of Environmental Biology Jornada Basin LTER Jornada Experimental Range - USDA School of Life Sciences - ASU Acknowledgments Lara G. Reichmann R.C.A. Gucho Owen P.B.R. McKenna Laura Yahdjian Deb Peters Kelsey McGurrin Josh Haussler John Angel III Shane & Miriam G. A. Gil Funding sources

45. Contrasting productivity responses to interannual precipitation variability Laureano A. Gherardi and Osvaldo E. Sala Arizona State University, School of Life Sciences Results of 6 years of precipitation manipulation at the Jornada Basin LTER

46. Precipitation Variability is projected to increase IPCC. 2013. Climate Change 2013: The physical science basis. Contribution of working group I to the fifth assessment report of the Intergovernmental Panel on Climate Change. Cambridge, United Kingdom and New York, NY, USA.

47. Interannual Precipitation Variability ANPP Mean ANPP CV Productivity Stability

48. NSF-Division of Environmental Biology Jornada Basin LTER Jornada Experimental Range - USDA School of Life Sciences - ASU Acknowledgments Lara G. Reichmann R.C.A. Gucho Owen P.B.R. McKenna Laura Yahdjian Deb Peters Kelsey McGurrin Josh Haussler John Angel III Shane & Miriam G. A. Gil Funding sources

49. ANPP, PPT, Space ANPP = - 45.13 + 0.67*MAP r 2 = 0.76 Bai et al (2008) ANPP=-34+0.60*MAP r 2 =0.94 Sala et al (1988) ANPP=-30+0.47*MAP McNaughton et al (1989) Great Plains South America Mongolian Plateau

50.  A simple model accounts for a large fraction of ANPP variability across space and for most grasslands of the world

51. 050010001500 0 200 400 600 800 1000 Annual Precipitation (mm) Temporal Model Spatial Model Aboveground Net Production (g/m 2 /yr) Lauenroth and Sala 1992 Ecological Applications 2:397-403 -34 + 0.60*MAP r 2 = 0.94, p < 0.001 56 + 0.13*MAP r 2 = 0.39, p < 0.001 Spatial vs. temporal models of net primary production

52. Sala et al 2012, Philosophical Transactions of the Royal Society B

53. r 2 =0.39 Sala et al 2012, Philosophical Transactions of the Royal Society B

54.  Time and Space cannot be exchanged for the ANPP-MAP relationship  Spatial model does not work through time  Temporal model only accounts for a small fraction of the variability explained by spatial models and has shallower slope

55.  Differences between spatial and temporal models are explained by time lags in ecosystem response to changes in water availability

56.  Time lags result from legacies of wet and dry years  ANPP observed = F (PPT t, Legacy)  Legacies = ANPP observed – ANPP expected ◦ ANPP expected = F (PPT t )  Magnitude of Legacy= F (PPT t-1 – PPT t )

57. Magnitude of Legacy= F (PPT t-1 – PPT t ) What is the shape of F ? How does this relationship change across a PPT gradient?

58. Sala et al 2012 PTRSB

60. Knapp and Smith (2001)

62. Sala et al 2012, PTRSB

68. Jornada LTER  MAP 240 mm  Dominant species: ◦ Bouteloua eriopoda C4 ◦ Prosopis glandulosa C3 Jornada Experimental Range Chihuahuan Desert Grassland

70. Fixed rainout shelters intercept different amounts of rain, depending on the number of shingles

71. Water was added to the increased PPT treatments after each PPT event, year around Total 132 plots

72. Reichmann, Sala, Peters, Ecology 2013 Legacy = -2.71 + 0.05 * ∆PPT R 2 = 0.42

73. Yahdjian and Sala (2002)

74. Precipitation input (mm/year) ANPP (g.m -2.yr -1 ) PPT mm/year 50 70 90 110 130 20 60 100140180220 without drought legacy after 80% rainfall interception after 55% rainfall interception after 30% rainfall interception Yahdjian and Sala (2006)

75.  Changes in precipitation result in legacies  Magnitude of Legacies is a function of difference in precipitation of current and previous year  Legacies in the Chihuahuan desert ecosystem are symmetrical ◦ │ Positive legacy │ = │ Negative legacy│

76.  Positive legacies would compensate negative legacies  Increased precipitation variability would not affect average productivity

77.  Structural mechanism ◦ Meristem density constrains production response to a wet year after a dry year ◦ Meristem density enhances production after wet years  Biogeochemical mechanism ◦ N limitation constrains production response to a wet year after a dry year ◦ Abundant reactive N enhances production after wet years  Soil moisture carry-over

78. Reichmann, Sala, Peters, Ecology 2013

80. (Reichmann and Sala, Functional Ecology 2014)

82. Biogeochemical mechanism Reichmann, Sala, Peters, Ecology 2013

84. Reichmann et al Ecosphere 2012

88. Tiller density determines magnitude of legacies Biogeochemical mechanisms do not determine legacies Soil water carry-over does not determine legacies

89. Collins et al 2011 Present Future + Precipitation Time Hypothetical pattern

90.  Observational  Short term manipulations Most studies are

91. Can we predict press effects of directional changes in precipitation amount and variability based upon our understanding of pulse responses?

92. Smith MD et al (2009)

93. a) Ecosystem response variables are proportional to precipitation Response variable = b 0 + b 1 *PPT Increased precipitation Ambient Decreased precipitation 0 + + Ecosystem response variable Time

94. b) Ecosystem response variables are proportional to changes in precipitation and to the time that the ecosystem has been exposed to the new condition Response variable = b 0 + b 1 *PPT + b 2 *Time Increased precipitation Ambient Decreased precipitation 0 + + Ecosystem response variable Time

95. c) Acclimation / exhausting of resources Response variable = b 0 + b 1 *PPT + b 2 *Time + b 3 *Time*PPT Increased precipitation Ambient Decreased precipitation 0 + + Ecosystem response variable Time

96.  The effect of time varies for different response variables Increased precipitation Ambient Decreased precipitation 0 + + Ecosystem response variable Time

97.  The effect of time is asymmetric for reduced and increased precipitation Increased precipitation Ambient Decreased precipitation 0 + + Ecosystem response variable Time

98. Can we predict press effects of directional changes in precipitation amount and variability based upon our understanding of pulse responses?

99. Individual Species Response + 80% - 80%

100. Species Richness + 80% Ambient - 80 %

101. Linear and non-linear ANPP responses to precipitation

102. 1.Linear and non-linear ANPP responses to annual precipitation Precipitation

103. Annual ANPP (g m 2 yr -1 ) Therefore, NULL precipitation variance effect on ANPP. Linear ANPP responses to precipitation result in negative effects of dry years equal to positive effects of wet years. Precipitation

104. Annual ANPP (g m 2 yr -1 ) Non-linear ANPP responses to precipitation result in different effects of dry and wet years. Therefore, POSITIVE precipitation variance effect on ANPP. Precipitation

105. Annual ANPP (g m 2 yr -1 ) Non-linear ANPP responses to precipitation result in different effects of dry and wet years. Therefore, NEGATIVE precipitation variance effect on ANPP. Precipitation

106. Interannual Precipitation Variability ANPP Mean ANPP CV Productivi ty Stability

107. Plant-functional types show different stability response to PPT variability Ecosystem stability results from the aggregated response of plant types

108. Functional diversity increases with PPT variability Changes in relative abundance support such effect How do we explain these responses?

109. 1.Inter-annual precipitation variability itself has a negative effect on ANPP 2.Non-linear responses and changes in soil water distribution explain such effect 3.Aggregated plant-functional type responses determine overall ecosystem response

110. 1.Interannual precipitation variability has a positive effect on ANPP CV 2.Contrasting plant-functional type responses result in relative abundance change and increased diversity 3.Aggregated response of plant-types determines ecosystem stability