Covering over 60,000 square feet, A Savage Garden, is a combination arboretum and botanical garden. The plants are on berms to allow better drainage and provide different areas of interest. Natural and gravel paths run between different areas of the garden providing easy access for visitors as well as for maintenance. Each berm area has valved irrigation and electrical power available for use.
Design Principles
- Limit water consumption
* use low ks plants
* use drip irrigation
* use Et based water
delivery
* extensive use of mulch
* soil moisture monitoring
- Limit maintenance
* proper plant selection
* natural hedges
* use natives where
appropriate
- Use Sustainable practices
* no chemical fertilizers
* horse manure used
where needed
* salvaged wood chip mulch
* use plants to attract
natural predators
* limit herbicide use to
last resort
Design: I developed the vision and basic layout for the garden with design consultation from two local designers: Micheal Glassman of HGTV fame and one other designer. Obtaining input from two different designers worked well. Micheal has a great eye for layout and the 'bones' of the garden. He provided the basis for the orientation and the central all'ee. The other designer is a plant enthusiast that has come up with great ideas and sources for plants; he also did the design of the botanical portion of the garden.
Layout: The entrance to the garden is an all'ee running west to east leading to a focal point area. The eastern half of the garden is the arboretum and the western half the botanical garden portion. The west side of the garden borders a driveway which is separated by a row of cotoneaster, griviena and callistemon. The east side of the garden is a grove of quercus lobata, q. agrifolia, and q. wislizenii native oaks that runs along a drainage swale.
Arboretum:
- aesculus californica
- araucaria bidwillii
- cedrus atlantica 'glauca'
- cercis occidentallis
- cotinus coggyygria 'Royal Purple'
- ginko biloba
- gleditsia triacanthos
- koelreuteria paniculata
- quercus suber
- quercus wislizenii
- quercus agrifolia
- quercus lobata
- sequoiadendron giganteum
pendulum
Botanical Garden:
- agave parryi huachuensis
- agave victoria reginae
- agave vilmoriniana
- arbutus marina
- arctostaphylos
- berberis t. 'rosy glow'
- berbis t. 'velvet cloak'
- brachychiton populneus
- buddleia davidii 'harlequin'
- buddleia davidii 'purple knight'
- butia capitata
- caesalpinia gilliesii
- ceanothus concha
- ceanothus 'tuxedo'
- chilopsis linearis 'burgundy'
- chilopsis linearis 'bubba'
- cistus x pulverulentus 'sunset'
- cupressus arizonica 'blue ice'
- dasylirion wheeleri
- euryops pectinatus
- feijoa sellowiana
- hesperaloe parviflora
- juniperus scopulorum 'wichita blue'
- langerstromia i. 'dynamite red'
- leucophyllum frutescens
- perovskia atripliafolia
- phlomis purpurea
- salvia apiana
- sambucus nigra 'black lace'
- vitex agnus-castus
Berms: Over 1800 cubic yards of soil was brought in to level the garden area and build planting berms . Berms are used to elevate plants for easier observation, provide separation from the walkways and improve drainage. Each berm has its own water supply, both valved and un-valved, and electricity.
Infrastructure: I designed and installed the irrigation system and electrical supply. Drip irrigation is used to precisely supply the optimal amount of water to each plant throughout the entire root system. The amount of water needed for each plant is calculated using evapotranspiration rates (Et) as determined from the listings at California Irrigation Management Information System (CIMIS). A 32 station irrigation controller is used which is programmed to change the water schedule according to the average monthly Et relative to Et max.
The garden also has high speed WiFi bridged into the garden. Power-line Ethernet is used for LAN connection on each berm.
A Few of Several Sprinkler Control Valves
Soil Moisture Monitoring: Soil moisture monitoring is implemented. The soil moisture is monitored at various locations of the garden to determine that the optimum amount of water is being delivered to the entire root zone. Sensors are located at two depths at each location: in the root zone to verify adequate moisture is available and below the root zone to detect over watering. The apparatus to collect and report the measurements was designed and built by myself. The system emails the soil moisture readings on a daily basis. A web page on my intranet provides a weeks worth of data. A PIC 8 bit microcontroller reads the watermark soil moisture sensors and sends the raw data to a raspberry pi for tension and content calculations, emailing of results and posting data to a web server.
Soil Moisture Measure System. A LCD Shows the Last Readings, Sensors are Connected to the Terminals
I built an irrigation monitoring system that monitors each irrigation station for water flow during irrigation, adherence to irrigation schedule and water flow outside of the irrigation schedule. The system can detect under or over watering, sprinkler valve faults, missed cycles and irrigation tubing leaks. An email is sent after each irrigation cycle indicating success or any detected deviations. An email notice is also sent should there be water flow outside of an irrigation cycle.
A calibrated water flow meter with adequate flow capacity is used to measure water flow. The flow meter provides 10 pulses per gallon. These pulses are counted and conditioned by a PIC microcontroller in an enclosure located near the flow meter. The microcontroller interfaces to an LED display in the garden that displays current water flow in gpm, amount of water that has flowed in the last hour and the last 24 hours. Having an infield readout is very handy during irrigation system modification and maintenance. The microcontroller also pulses, using a solid state relay, a current loop to the main controller located 150 feet away in the irrigation control box. The pulses are condition pulses from the water flow meter.
In the irrigation control box the main controller, a Raspberry Pi 2, is housed. The Rpi is interfaced to boards that sense sprinkler valve solenoid current, counts pulses on the current loop from the water flow microcontroller system and has a LAN connection for internet access. Python programming implements the system to measure water flow through each individual valve, self calibration of water flow, comparison to irrigation schedule, detecting water events outside of the irrigation schedule, saving and posting data and sending email alerts.
Comparison to the irrigation schedule is able to comprehend a schedule that varies monthly, incorporates delays for well replenishment and multiple cycles per day. The data is transferred to text files for non-volatile storage and access for web server access.
Each valve current sensor board is able to monitor 8 valves with a maximum of three daisy chained boards for 24 valves total. In the Savage Garden only 14 valves are used so only two current sensor boards were built. Each valve's current is monitored independently by measuring voltage drop across a sense resistor. Since sprinkler valves use 24 Vac the voltage has to be rectified, filtered to remove noise and then interfaced to the Rpi input through an opto-isolater. The opto-isolater further reduces noise and protects the Rpi's less than robust inputs. A push button on the board is used to start water flow calibration. If modifications to an irrigation circuit has been made, such as adding emitters, the system can be re-calibrated to the new flow rate by simply running that irrigation circuit and pushing the button.
By logging into the Rpi using SSH, the irrigation schedule and water flows are inputted as well as reporting parameters.
The irrigation monitor has already demonstrated its usefulness in detecting irrigation tubing leaks and solenoid valve failures immediately after occurrence. This prevented plant damage and excessive water use.
Irrigation Monitoring System Main Controller and Valve Current Sensing Inside Irrigation Controller