Spatial and Temporal Controls of Aboveground Net Primary Production

Transcription

1 Spatial and Temporal Controls of Aboveground Net Primary Production PRODUCTIVITY, WATER, SPACE AND TIME JORNADA COURSE JULY 2013 Osvaldo Sala

2 ANPP, PPT, Space Sala et al (1988) Great Plains ANPP= *MAP r 2 =0.94 South America McNaughton et al (1989) ANPP= *MAP Mongolian Plateau Bai et al (2008) ANPP = *MAP r 2 = 0.76

3 Sala et al 2012, PTRSB

4 Message A simple model accounts for a large fraction of ANPP variability across space and for most grasslands of the world

5 Spatial vs. temporal models of net primary production Aboveground Net Production (g/m 2 /yr) Temporal Model Spatial Model *MAP r 2 = 0.39, p < Annual Precipitation (mm) *MAP r 2 = 0.94, p < Lauenroth and Sala 1992 Ecological Applications 2:

6 Sala et al 2012, PTRSB

7 Spatial vs. Temporal models of net primary production r 2 =0.39 From Sala et al 2012, PTRSB

8 Message Time and Space cannot be exchanged for the ANPP-MAP relationship Spatial model does not work through time Temporal models only account for a small fraction of the variability explained by spatial models and have shallower slope

9 Hypothesis Differences between spatial and temporal models are explained by time lags in ecosystem response to changes in water availability

10 Legacies 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 )

11 Global patterns of Legacies Magnitude of Legacy= F (PPT t-1 PPT t ) What is the shape of F? How does this relationship change across a PPT gradient?

12 Legacy Symmetry Hypotheses Sala et al 2012 PTRSB

13 Legacy Symmetry Hypotheses

14 H 3.2 Knapp and Smith (2001)

15

16 Effect of previous-year PPT on ANPP across sites Sala et al 2012, PTRSB

17 Effect of current year PPT on ANPP Sala et al 2012, PTRSB

18 Effect of previous-year PPT on ANPP across sites Sala et al 2012, PTRSB

19 Effect of previous-year ANPP on current-year ANPP across sites Sala et al 2012, PTRSB

20 Effect of current- and previous-year PPT along a PPT gradient Sala et al 2012, PTRSB

21 Experimental Approach Chihuahuan Desert Grassland

22 MAP 240 mm Dominant species: Bouteloua eriopoda C4 Prosopis glandulosa C3 Jornada LTER

23 Experimental design

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

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

26 Legacy Magnitude Legacy = * PPT R 2 = 0.42 Reichmann, Sala, Peters, Ecology 2013

27 Rainout shelters in the Patagonian steppe Yahdjian and Sala (2002)

28 Drought legacies in the Patagonian Steppe ANPP (g.m -2.yr -1 ) without drought legacy after 80% rainfall interception after 55% rainfall interception after 30% rainfall interception 50 Precipitation input (mm/year) PPT mm/year Yahdjian and Sala (2006)

29 Conclusion 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

30 Corollary Positive legacies would compensate negative legacies Increased precipitation variability would not affect average productivity

31 Hypotheses for the Legacy Mechanisms 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

32 Structural mechanism Reichmann, Sala, Peters, Ecology 2013

33 Structural mechanism

34 Biogeochemical mechanism

35 Biogeochemical mechanism

36 Biogeochemical mechanism

37 Test of the soil-moisture carry-over hypothesis

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

39 Ecosystem Response to Long-term Changes in Precipitation

40 Pulse and Press External drivers Climate, globalization Social template Human behavior Policy, markets, reproduction and migration H5 H6 Pulses: fire, drought, storms, dust events, pulse nutrient inputs, fertilization Presses: climate change, nutrient loading, sea-level rise, increased human resource consumption H1 Biophysical template Community structure Species turnover time, trophic structure, microbial diversity H2 Human outcomes Quality of life, human health, perception and value H4 Ecosystem services Regulating: nutrient filtration, nutrient retention, C sequestration, disease regulation, pest suppression Provisioning: food, fiber, and fuel Cultural: aesthetics and recreation Supporting: primary production, nutrient cycling Ecosystem function Flux, transport, storage, transformation, stoichiometry, primary productivity H3 Collins et al 2011

41 Press Conceptual Model Smith MD et al (2009)

42 Hypothesis 1 a) Ecosystem response variables are proportional to water availability Ecosystem response variable + Time 0 + increased water ambient decreased water Response variable = b 0 + b 1 *PPT

43 Hypothesis 1 b) Ecosystem response variables are proportional to changes in water and to the time that the ecosystem has been exposed to the new condition Ecosystem response variable Time + increased water ambient decreased water Response variable = b 0 + b 1 *PPT + b 2 *Time

44 Multiyear PPT trend precipitation (cm) (a) summer spring dormant season + b a water year (Oct-Sep) + Peters et al (2011)

45 Consequences of Multiyear PPT trend 500 (a) Upland grasslands ANPP (g/m 2 ) time since 1993 (year) Peters et al (2011)

46 Hypothesis 1 c) Acclimation / exhausting of resources Ecosystem response variable Time + increased water ambient decreased water Response variable = b 0 + b 1 *PPT + b 2 *Time + b 3 *Time*PPT

47 Hypothesis 2 The effect of time is asymmetric for reduced water and increased water Ecosystem response variable Time + increased water ambient decreased water

48 Hypothesis 3 The effect of time varies for different response variables Ecosystem response variable Time + response variable A response variable B

49 ARMS automated rainfall manipulation system Solar panel Intermediary tank (55 gal.) Battery Interception plot Filter Float switch Pump Irrigation plot Gherardi & Sala, Ecosphere 2013

50

51 Conclusions 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

52 Thank you Laureano Gherardi Lara Reichmann Owen McKenna Josh Haussler Kelsey Duffy

53

54 Questions How do magnitude and symmetry of legacies affect the ecosystem response to changes in variability? Why the effect of ppt on ANPP decreases with MAP? Why plants do not deploy the optimal density of tillers each year to maximize ANPP and resource use? Use Schwinning Sala graph of two classes ago Why there are no legacies associated with microbial mediated processes?