Perfection the Enemy of Progress...

**Drfig** · 07-23-2021, 12:30 PM

DUDE, your like mad scientist! its awesome.
dont forget the 1.21 jigawatts

**Figgerlickinggood** · 07-23-2021, 12:49 PM

Isn’t there some way to write this all out in layman’s term?

**Ktrain** · 07-23-2021, 01:23 PM

Excellent, thank you for sharing this is amazing.

**ieatfigs** · 07-23-2021, 01:34 PM

Nice job explaining GDD. I could never understand what that means, until now.

**davej** · 07-23-2021, 02:02 PM

I know there has been some discussion here in the past regarding GDD's compared to days above 50F. The thought is that any day over 50F is approximately equivalent. Obviously the GDD approach presumes that a fig will develop and ripen much better at 110F than it will at 60F but does the evidence actually bear this out?

**VentSolaire** · 07-23-2021, 02:16 PM

I am a new gardening hobbyest so I apologize in advance if what I am going to suggest has been discussed before.

Originally posted by AscPete View Post

IMO the added complexity would typically only change most locations' totaled seasonal GDD50 Value an equivalent of less than one to two calendar week(s) maximum.

My ripening window is very short: about a month, typically from early September to early October. A variation of 1-2 calendar weeks is 25-50% of my window, it is not negligeable. Increased accuracy in predictability is desirable.

All models are wrong, but some are useful.

I am sure there are many models developped in agriculture for growth and harvest calculation. One component of the GDD model that you have mentioned is to put a base value (e.g. 50F), another one would be to use also a cap value (80F).

In mathematical form:

However by doing so, the weight for each degree is the same. Each additional degree leads to one additional GDD50, no matter if it's going from 50F to 51F, 65F to 66F, or 89F to 90F.

Are you aware if sigmoid functions (e.g. the logistic function, or the cumulative distribution function for the normal distribution) have been explored in the past to assign different weights for each additional degree? Sigmoid functions have a parameter to adjust how smooth you want the increase in the lower and high end.

I'm thinking something like:

or

Graphically all three models look like this:

All three models are capped at 50F and 80F. The sigmoid function has the additional benefit that it would address the issue that a grower at 70F accumulate GDD much faster than someone at 60F, not just by a factor of 2 (as with GDD50) but 5.2 (using the erfc sigmoid function).

**Miracle fruit** · 07-23-2021, 03:09 PM

Great info Pete. It shows your dedication to growing figs, this write ✍️ is so Nobel of you.
Thanks for sharing 👏🤙

**batty** · 07-23-2021, 03:21 PM

Originally posted by AscPete View Post

IMO the added complexity would typically only change most locations' totaled seasonal GDD50 Value an equivalent of less than one to two calendar week(s) maximum.

I had meant to respond to your similar thought on our last convo, but got busy! I had a hard time seeing this so I investigated.. I looked for extreme locations, and completely failed to find a location where a cap would make a significant difference in the prediction - hot areas tend to also have long springs, so impact of those really hot summer months is minimal for ripening prediction. Truly tropical areas would be different but for most temperate locations the difference is < 1 week difference for a fig ripening target GDD of 2-3k, even if using a more aggressive cap at 86F (vs no cap). The USpest site is nice if anyone would like to compare models. Takeaway: no real need to add the temperature cap, except in extreme situations.

Are you suggesting the spreadsheet (or new data collection) be re-opened for additional data?

I think it would be helpful to collect more data in a central spot to help the lack of data on some varieties, but also to build confidence in the model and get newer folks more involved.

**WIFigger** · 07-23-2021, 07:11 PM

Thanks for posting. I haven't read it all yet but I wanted to share this link for those of us in the Midwest. It's built for corn but can be adapted.

https://mrcc.illinois.edu/U2U/gdd/

**jrdewhirst** · 07-23-2021, 07:15 PM

Pete --

Your title ("Perfection the Enemy of Progress") may be obscure to any reader who did not catch our prior discussion. In that discussion, I pointed out some shortcomings of the GDD measure. You defended GDD50 as a good enough measure of cumulative heat, useful for predicting fig ripening. I agreed with you that GDD50 is useful, easily calculated, and close enough to the truth of fig physiology to work for most growers in most settings. I made it very clear that I would not want my pursuit of improvement to dissuade people from using GDD50 as a "good enough" tool. As I said, I did not want "the perfect to be the enemy of the good." I think your title refers to my comment implicitly. In effect, you are saying, "Don't bother trying to modify the heat measure in order to improve our predictions of fig performance."

https://en.wikipedia.org/wiki/Perfec...enemy_of_good: "Achieving absolute perfection may be impossible and so, as increasing effort results in diminishing returns, further activity becomes increasingly inefficient."

Your title seems to imply that pursuing improvement impedes progress. I don't see how. There's no reason why most growers can't continue to use GDD50, if it meets their needs. Others (e.g., me) can waste our time exploring alternative measures.

Why should we?

Well, you yourself make clear that the linear GDD function is wrong. As you write above, " > 50°F (10°C) grow increases to ~ 77°F and > 85°F growth slows." This summary of fig physiology is best represented by a non-linear sigmoid function such as the one presented above by VentSolaire.

Your fall-back defense of GDD50 is that the more accurate sigmoid function is only immaterially better. But whether GDD50 is good enough seems to depend on growing conditions. ft the truth looks something like VentSolaire's sigmoid function, which is fairly consistent with your own description of how figs grow, then GDD50 has the potential to mislead in 3 situations: The GDD50 measure will accumulate (1) too many GDD when temperatures are ~50-65 F, (2) too few GDD when temperatures are ~65-80 F, and again (3) too many GDD when temperatures are >80 F. Basically, growers with either cool conditions or hot conditions will be too optimistic; growers with merely warm conditions won't be optimistic enough. Whether these errors are big enough to matter should be up to the grower to decide.

This criticism is not merely hypothetical. The GDD50 measure can mislead growers by suggesting that temperatures over 80 F are helpful for fig growth and ripening. Uncapped GDD50 says, "Hotter is better." In fact, my digression into measures of heat occurred within another thread where a grower there was attributing the great performance of his trees to high heat. You and I agree that he was wrong. But that's the kind of misconception GDD50 leads to. "I gotta get my GDDs up -- Hotter is better."

This would seem a trivial one-off EXCEPT that there must be thousands of forum members imagining that they'd get more figs faster if they could just build a big greenhouse and let it heat to 100 F. But as you write above, ">85 F growth slows." That would be obvious with a capped GDD measure.

OR . . . the GDD measure can be improved along exactly the lines described above by VentSolaire. The calculation is actually simpler than the standard GDD50.

**WIFigger** · 07-23-2021, 09:05 PM

Do you think the extra hour or 2 of daylight in Northern US areas compared to Southern have any impact on ripening?

**straywolf94** · 07-23-2021, 10:39 PM

Thank you for simplifying and explaining this AscPete. Can we pin this to frequent topics as a reference?

**DrDraconian** · 07-24-2021, 12:54 AM

My take on this:

From May through September (the primary fig production months), Chicago is actually a better place to grow and ripen figs than San Diego. And that makes me just a little bit sad

(although I'm not at all surprised)

**AscPete** · 07-24-2021, 03:31 AM

DrDraconian ,

Sorry for your Lows...

Its clear from the Monthly GDD50 values that you're always at the lower end. The growing season's monthly GDD50 Values are typical of mid - late Spring in the North East.
Is sunlight adequate in late winter, February and March?

**WVMJack** · 07-24-2021, 05:27 AM

Maybe we could translate this into a map of the USA that shows areas that could say ripen 50% of the figs on a Black Madeira with no heroic intervention? Maybe different shaded areas for the latest commonly grown late figs that can ripen in that zone? The 50% standard would be good for ripening, as some people ripen 1 black madeira on their tree and the rest never get ripe, that date is not a useful date when you only get 1 fig vs %50.

**Citicodays** · 07-24-2021, 06:47 AM

AscPete
So I have over 4000 GDD between mid April and mid October, when average lows are above 50 degrees. This equates to a 6 month season in Knoxville TN. I am willing to protect from late frosts. Do you interpret this to mean that I should be able to ripen any late fig on an average year?

**jrdewhirst** · 07-24-2021, 08:57 AM

Pete's data on Richard's climate illustrate the issue. Pete calculates that Richard has 4351 GDD. All good, right?

But most of the days Feb-April struggle to get over 65 F. That's almost 1000 GDD of questionable real value. If the truth of fig physiology is anything like the sigmoid function displayed by VentSolaire in Post #7, then Richard's fig season kicks off in May (like mine).

If GDD50 works as advertised (see the OP), then the Mt Etnas ripen after 2200 degree days, which in San Diego is mid-July.

DrDraconian -- Can you tell us when a few benchmark varieties actually ripen for you?

**AscPete** · 07-24-2021, 09:56 AM

Originally posted by Citicodays View Post

So I have over 4000 GDD between mid April and mid October, when average lows are above 50 degrees. This equates to a 6 month season in Knoxville TN. I am willing to protect from late frosts. Do you interpret this to mean that I should be able to ripen any late fig on an average year?

Although your Accumulated GDD50 is over 4000° Days (4048 Degree Days, mid-April to mid-October) the monthly totals on the overlay below shows a slightly different picture, but the numbers look good for late figs.
IMO, It would be better to get an early start in April (with protection) and "Ripen" the figs in July and August, Low Tunnel?

... https://www.timeanddate.com/weather/...xville/climate

GDD50 values are calculated from the monthly Average Temperatures.
By July 1, its either 1323°Days (cold April) or 1608°Days (GDD50)
By September 1, its either 3093°Days (cold April) or 3378°Days (GDD50)
And like most areas the weather in September (Season End) Cools, reducing available GDD's.

**AscPete** · 07-24-2021, 11:42 AM

Originally posted by jrdewhirst View Post

Pete -- Sometimes you tear down straw men that only vaguely resemble what I said, giving the impression that you addressed a point when actually you didn't.

1. << IMO, having a seasonal GDD50 Value that's one or even two weeks off, or ~ 5% should not be cause to delay its implementation for a more "Accurate Solution". >>
I never suggested that we delay the implementation of GDD50. I suggested that (a) we use GDD50 while looking for something better; and (b) we then switch if the benefits seem meaningful in our growing environment.
Also as others noted, in a cool growing region an error of 1-2 weeks is as much as 25% of the growing season. A 5% error in a measure of heat accumulated over 5 months can translate into a big crop of unripe figs.

2. << And Yes, Greenhouses at 100°F days with 70° - 80°F nights will produce much more growth than 50s - 60's Day / Night... >>
I never suggested that a greenhouse wouldn't help someone whose temperatures otherwise would hover ~50-60 F. I suggested that a 100 F greenhouse would be useless (or worse) for a grower whose daytime summer temperatures are already 65-80 F. An uncapped GDD gives the impression that hotter is ALWAYS better.

3. << The "Accurate Solution" could always be implemented as an upgrade at a later date, just as GDD50 was "upgraded" to GDD 50/85. >>
Yes, this is exactly what I'm suggesting. The simplest upgrade would be to use GDD incorporating a cap near 80 F. A more sophisticated upgrade, but not computationally more difficult, would be to use a sigmoidal function such as described above.

1. As we have previously discussed the error would be on the "Cap" / High Limit, which is covered by the max 30°day / 900°month. The Low limit is covered by the Base50°F (Base10°C) which was not in your initial calculations, since you were looking far a "Warm" Low Limit Value. The High Limit / "Cap" would only introduce a nominal error (< 1%), in any Temperate Zones that experience any days over the 30°day GDD50 maximum... The High Limit / Cap will eliminate reduce the error for Hot Zones as shown in the diagram of Las Vegas in the OP, the nominal 5% is a conservative estimate, the actual calculations were < 3%.

2. Daytime temperatures between 65 - 80 degrees would definitely be benefited by a Greenhouse, the optimal Growth Temperatures for Ficus carica L appear to be between ~ 77° - 86°F (25° - 30°C),and as mentioned in the OP the High Limit / Cap is 30°day GDD50..

3. As mentioned in the previous topic the 80°F (26°C) High Limit Cap would be too low, when you account for the daily Ambient Temperature swings from Low to High (sawteeth), it would introduce a much greater error / deviation...

Originally posted by jrdewhirst View Post

Pete's data on Richard's climate illustrate the issue. Pete calculates that Richard has 4351 GDD. All good, right?
But most of the days Feb-April struggle to get over 65 F. That's almost 1000 GDD of questionable real value. If the truth of fig physiology is anything like the sigmoid function displayed by VentSolaire in Post #7, then Richard's fig season kicks off in May (like mine).
If GDD50 works as advertised (see the OP), then the Mt Etnas ripen after 2200 degree days, which in San Diego is mid-July.
DrDraconian -- Can you tell us when a few benchmark varieties actually ripen for you?

The "sigmoid function" could possibly be more accurate but where do we get the required data to "assign different weights for each additional degree"?
My argument has always been to use the currently available info to get the Forum Community informed about and comfortable using the simple GDD50 Metrics.

Richard's experiences in San Diego is easier to understand when the monthly GDD50 Values are examined compared to Fig Culture.
He has moderate to low GDD50 (Heat), hopefully there is some compensation from the Solar Gains and increased DLI (Daily Light Integral).

GDD50 values are calculated from the monthly Average Temperatures
The monthly GDD50 values clearly show "Low Growth" January - April (25% - 40%)
"Moderate Growth" May - August (50% - 74%)
Declining to "Low Growth" September - December (69% - 24%)

**WVMJack** · 07-24-2021, 02:15 PM

I put a couple of data loggers on my figs this year, hoping they keep recording all the way until Nov when leaf fall will be the time to pull them off and see what the actual GDD are, also recorded when figs set and plan to record when they ripen just for fun. I hope to compare it to the predicted one from one of the GDD calculators.

**AscPete** · 07-24-2021, 03:02 PM

Originally posted by Figgerlickinggood View Post

Isn’t there some way to write this all out in layman’s term?

**jrdewhirst** · 07-24-2021, 08:28 PM

AscPete -- Just a quick response on a few points.

<< The "sigmoid function" could possibly be more accurate but where do we get the required data to "assign different weights for each additional degree"? >>

Where do we get the weights for GDD? We make them up! Every degree of temperature above the floor gets the same identical weight. But that's totally arbitrary. It only avoids seeming arbitrary because we have an unconscious bias that anything equal is natural and fair.

In this case, I'd suggest two possibilities:

First, you could translate your guidance into numbers. You wrote:

The Published Academic Horticultural Research is that the temperature limits are approximately;
< 45°F growth stops and Dormancy @ < 40°F (4°C).
> 50°F (10°C) growth increases to ~ 77°F and
> 85°F growth slows

So make a chart where the X axis is Temperature and the Y axis is Growth. Draw a sigmoid-ish curve that fits your guidance (above) and/or experience. Use the values from your own guesstimated curve.

Or second, we could pick a function that seems to have the right shape and take the values it produces. The ERFC curve presented in Post #7 looks plausible. If you don't like it as presented, you could shift the midpoint and change the range. In the example, the midpoint is 65 F and the range is 30 F degrees (+/- 15 F). So there is a floor (zero growth) at 50 F and a cap (maximum growth) at 80 F. This may not be exactly right, but hell - you already told us not to aim for perfection. And its is obviously better than GDD.

<< My argument has always been to use the currently available info to get the Forum Community informed about and comfortable using the simple GDD50 Metrics. >>

I agree with this approach. For the majority of forum members, GDD50 is good enough. But members should be warned that it will give misleading results if the growing environment is consistently cool (50-65 F) or hot (>80 F). Growers who have the interest to pursue other measures should be pointed in the right direction.

<< Richard's experiences in San Diego is easier to understand when the monthly GDD50 Values are examined compared to Fig Culture. He has moderate to low GDD50 (Heat) >>

Yes, exactly. His growing environment is cool. So it takes him way longer to ripen figs than GDD suggests. But once we understand Richard's experience, can't we do something to help him out? Can we give him a heat metric that will better predict when his figs will ripen? I think so.

**Citicodays** · 07-24-2021, 08:59 PM

AscPete
Here’s my Smith breaking bud in late March with CAREFUL use of a black garbage can. Actually worked out very well.

**ZomVee** · 07-24-2021, 09:16 PM

Some of you are like:

Meanwhile, I'm like:

Perfection the Enemy of Progress...

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment