Quantitative Approaches to Riparian Restoration in California (USA)
John C. Stella Dept. of Environmental Science, Policy and Management University of California, Berkeley and Stillwater Sciences [email protected] Restauración de Ríos Seminario Internacional Madrid, 20 Septiembre 2006
Outline
1. Riparian forests in California’s Mediterranean climate zone
2. Historical human impacts to the ecosystem 3. Deciding what to restore--processes or structure? 4. Quantitative approaches to restoring riparian forests -restoring ecological processes efficiently -restoring riparian structure effectively
1 Non-Equilibrium Ecosystems: Multiple Disturbances and Drivers of Change Fire Floods
Climate change
Landscape modification
Sacramento River Length: 615 km Basin area: 70,000 km2) Sacramento River Basin
San Joaquin River San Length: 530 km Francisco Basin area: 83,000 km2 Major tributaries: Tuolumne, Merced, Stanislaus Rivers
Major California River Systems
California Department of Water Resources.
2 Riparian Structure and Pattern
Herbaceous cover Cottonwood forest Mixed riparian forest Valley oak forest • High structural complexity
• Patchy distribution
• Important terrestrial and in-stream habitat (litter, large woody debris, shade)
Riparian Vegetation Establishment Processes on Alluvial Rivers
RiverRiver channel channel
TerraceTerrace FloodplainFloodplain
PointPoint bar bar
Channel Increasing age migration of vegetation
Floodplain Terrace Eroding River Point bar (poplar/willow (valley oak bank channel (gravel & scrub) mixed forest) woodland)
3 Human Impacts in Riparian Zones
Floodplain development
Habitat fragmentation
Channelization and bank stabilization
Water Development in California’s Central Valley
Major Water Shasta Dam, Projects Sacramento (CA Dept of Water River Resources)
California Aqueduct
4 50000
45000
40000 ExchequerExchequer Dam Flow Regime constructed, 1926
35000 Dam, 1926 NewNew Exchequer Dam Modifications constructed, 1967 30000 Dam, 1967
25000
Discharge (cfs) 20000 Reduction in annual 15000 flood magnitude 10000
5000
0 1900 1905 1910 1920 1925 1930 1935 1940 1945 1950 1955 1960 1965 1970 1975 1980 1985 1990 1996 Year 600 riparian tree establishment
/s) 500 3 Reduced flow 400 volume and 300 unimpaired
altered timing discharge (m 200 regulated
100
0 1 1 1 1 1 1 1 1 1 n b ar ay n l p Ja Fe M Apr M Ju Ju Aug Se
Floodplain Destruction: Gold Dredging, 1900-1950
Merced River Floodplain
Vegetated depressions <30 m
Image courtesy of Stephen Johnson
5 Ecological Impact: • few remnant forest patches • no seedling establishment • conversion to other vegetation types
What Should We Restore in a Disturbance-Dependent Ecosystem?
1. Ecological processes:
(e.g., flow variability, sediment transport, channel migration)
2. Habitat structure and pattern
(e.g., species composition, canopy configuration, age structure)
6 General Restoration Approach • understand hydrologic dependencies of riparian and aquatic species • model hydrology and stratify floodplain by degree of hydrological connectivity • choose meaningful quantitative metrics for each restoration approach • adapt analyses from other disciplines as necessary
San Joaquin River Hydraulic and Topographic Model
Seedling Recruitment Processes for Riparian Populus and Salix
• Populations limited by seedling recruitment
• Short-lived seeds; no seed bank
• Reproduction timing coincides with regular spring floods
• Seedlings establish in high-flow years
7 Critical Ecological Processes for Seedling Establishment •river flow regime • seed release timing • seedling water stress thresholds
’ Maximum discharge river stage Banks dry during seed release
ABOVE BASEFLOW ABOVE Water stress TION VA
ELE Successful recruitment Mahoney JM, and Rood SB. 1998. Summer baseflow APR MAY JUN JUL AUG SEP Wetlands 18:634-645.
Seed Release 20 Fremont cottonwood 2002 Goodding's black willow 120 16 Narrow-leaved willow Timing and unimpaired discharge 12 80 Spring Floods 8 40 4 0 0 20 2003 120 16 • Peak seed release 12 80 coincides with 8 40 snowmelt recession 4 0 0 • Annual timing 20 2004 120 influenced by 16 unimpaired discharge (1000 cfs) temperature 12 80 seed release index (open catkins/tree) 8 40 4 0 0 1 1 1 1 1 1 1 1 ar r y n l g p ct M Ap Ma Ju Ju Au Se O
8 Modeling Seed Release Timing From Seasonal Heat Sums Fremont cottonwood Goodding's black willow Narrow-leaved willow 120 140 160 180 observedseed release (day of year)
120 140 160 180 predicted seed release (day of year)
Seedling Water Stress Experiment • seedlings grown in tanks • 5 water table reduction treatments • measured seedling survival, growth, and physiology
9 100 Modeling Cohort Survival 80
60 from Empirical Data
40 drawdown rate cohort survival (%) 0 cm/d 1 cm/d 20 3 cm/d 6 cm/d 9 cm/d 0 0 1020304050100 days since start of drawdown 80 60 40 20 0
Flow Requirements for Successful Recruitment
600
Populus seed release 500 Salix seed release /s)
3 400
300
unimpaired 200 regulated discharge (m proposed
100
0
1 1 1 1 1 1 n b 1 ar r 1 y n l 1 g p Ja Fe M Ap Ma Ju Ju Au Se
10 Restoring Riparian Habitat Structure on Dredger Tailings
dredger slough with patches of riparian vegetation river channel
f lo w dredger tailing piles
Merced River Riparian Tree Planting Experiment
• 4 native species • 3 elevation levels • irrigation (+/-) • survival and growth analysis
11 Predicting Survival in Year 1 from Initial Planting Size (Logistic Regression Models)
POFR
1
8
.
0
6
.
0 4
. 80% chance of 0
2 survival for stems .
0 >20 mm diameter
survival probability
0
60 50 in 40 30 tia 5 l 2 he 30 0 ig 2 m) ht 2 5 r (m (c 0 1 ete m 10 iam ) 1 al d 0 5 initi
Species Survival Differences and Irrigation Effectiveness (Cox Proportional Hazard Models)
Irrigated both years Unirrigated in Year 2
Box elder Fremont cottonwood 0.0 0.2 0.4 0.6 0.8 1.0 0.0 0.2 0.4 0.6 0.8 1.0
0 20406080 0 20406080 cohort survival cohort
Oregon ash Valley oak 0.0 0.2 0.4 0.6 0.8 1.0 0.0 0.2 0.4 0.6 0.8 1.0
0 20406080 0 20406080 weeks since start of experiment
12 Conclusions
1. Restoring non-equilibrium ecosystems requires novel approaches.
2. Choose what to restore: processes and/or habitat structure.
3. Restoring processes efficiently requires understanding key ecological mechanisms.
4. Restoring habitat effectively requires controlled studies to isolate treatment effects from covariate factors.
5. Developing meaningful quantitative metrics is critical.
Thank you for your attention!
Special thanks: Marta González del Tánago Ministerio de Medio Ambiente
Collaborators: John Battles, Dept of ESPM, UC Berkeley Joe McBride, Dept of ESPM, UC Berkeley Matt Kondolf, Dept of LAEP, UC Berkeley Stillwater Sciences, Berkeley CA
Funding: CALFED Bay Delta Program National Science Foundation University of California, Berkeley
13 Climate change…. a wildcard
Expected in California: • More precip falling in winter (rain and rain-on-snow events) •Earlier snowmelt snowmelt initiation (day of year) 20 40 60 80 100 120
1920 1940 1960 1980 2000 Question: How much can trees adapt?
14 Pre-settlement Reference State Terrace Valley oak and Bare alluvial bar and Riparian scrub and Cottonwood and mixed forest Valley oak forest grassland mixed forest sparse seedlings mixed willow (mid-successional) (late-successional) (late-successional) (very early-successional) (early-successional)
Extreme flood
Moderate flood Bankfull flow
Baseflow
Low flow channel
Overflow Active channel channel
Terrace Floodplain Bankfull channel Floodplain Terrace
Current Conditions
Orchards Senescent valley oak Channel with encroachedSenescent cottonwoodRow crops Orchards forest mixed forest forest
Extreme flood Moderate flood Bankfull flow Baseflow
Low flow channel
Active and OverflowAgricultural Bankfull channel channel levee
Former floodplain/ Terrace current terrace Current floodplain boundary Former floodplain/current terrace Terrace
15 California rivers and water projects Maps courtesy of California Department of Water Resources.
Flow Timing and Magnitude
250 Riparian tree recruitment Unregulated flow period )
-1 200 Regulated flow s 3
150
100
mean monthly discharge (m discharge mean monthly 50
0 OCT NOV DEC JAN FEB MAR APR MAY JUN JUL AUG SEP
16 Central Valley Water Development
unimpaired regulated
600 seedling recruitment period
/s) 500 3 Friant Dam (San Joaquin River) 400
300
discharge (m 200
100
0 1 1 1 1 1 1 1 1 1 n b ar ay n l p Ja Fe M Apr M Ju Ju Aug Se
Biologically-important flow characteristics
600 Winter Peak Flows (seedbed preparation) Spring Snowmelt (germination & 500 early growth) /s) 3
400
300
200 discharge (m
Summer Baseflow (year 1 survival) 100
0 1 1 1 1 1 1 1 1 1 n b ar y n l p Ja Fe M Apr Ma Ju Ju Aug Se
17 Conceptual Model of Key Ecological Processes
Seed release timing Seedling water stress thresholds
‘Recruitment’ box’ Maximum discharge river stage Banks dry during seed release ABOVE BASEFLOW ABOVE Water stress Suitable Site ELEVATION Successful hydrology recruitment Summer baseflow APR MAY JUN JUL AUG SEP
Adapted from: Mahoney JM, and Rood SB. 1998. Streamflow requirements for cottonwood seedling recruitment--an integrative model. Wetlands 18:634-645.
Using Degree-Days to Predict Timing • Degree-days are an alternative way to mark development time • Established method in agriculture and pest management • Experimental vs. empirical models
• Daily degree-days: Dd = Tm – θ ;
• Cumulative degree-day threshold: D* = ∑°Dd
30 Daily Temperature
15 Xθ (base temp) degrees C 0 500 Cumulative Degree-Days
279D* (threshold)D*
0
Apr 1 May 1 Jun 1 Jul 1 Observed development stage
18 Climate change…. a wildcard
Expected in California: • More precip falling in winter (rain and rain-on-snow events) •Earlier snowmelt snowmelt initiation (day of year) 20 40 60 80 100 120
1920 1940 1960 1980 2000 Question: How much can trees adapt?
19